You Can See Everything
← all movies. Each comment's label under every pathway (+ positive, − negative, 0 neutral / mixed / no verdict; hover a cell for the verdict).
Pick a pathway to show its verdict and rationale under every comment: VADER (old v1) · siebert+opinion (LIVE) · Local NLI stance (deberta-v3-large) · GPT-6.1 Sol · Claude Opus 5.5 · Claude Sonnet 5.5 · Gemini 4 Argon · Grok 4.7 · GPT-6 Luna (cheap) · Claude Haiku 5.5 (cheap) · Gemini 3.5 Flash-Lite (cheap)
| comment (excerpt ≤280 chars) | VADER (old v1) | siebert+opinion (LIVE) | Local NLI stance (deberta-v3-large) | GPT-6.1 Sol | Claude Opus 5.5 | Claude Sonnet 5.5 | Gemini 4 Argon | Grok 4.7 | GPT-6 Luna (cheap) | Claude Haiku 5.5 (cheap) | Gemini 3.5 Flash-Lite (cheap) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| rank · score · bucket · Reddit · comments +/0/− | #3 81 Watch R 66 58/18/27 | #7 71 Depends R 47 33/36/34 | #8 56 Depends R 20 11/62/30 | #1 95 Watch R 93 35/67/1 | #1 95 Watch R 93 39/63/1 | #1 93 Watch R 89 30/71/2 | #1 95 Watch R 93 37/65/1 | #2 92 Watch R 88 39/60/4 | #4 89 Watch R 82 39/57/7 | #4 89 Watch R 81 33/65/5 | #1 93 Watch R 89 33/67/3 |
| thread #0 w 1.0 “I know how you’re feeling right now” “No you don’t” Got the biggest laugh out of my crowd. Just a cathartic line after listening to that douche bag ramble on for an hour. VADER (old v1): positive — VADER 0.382 siebert+opinion (LIVE): positive — P(op)=0.49 P(pos)=0.99 Local NLI stance (deberta-v3-large): negative — NLI p=0.41 GPT-6.1 Sol: positive — Praises the line as cathartic and highlights its successful comedic impact on the audience. Claude Opus 5.5: no_verdict — Quotes a line and describes the crowd's reaction; calling a character a douchebag isn't a judgement of the film. Claude Sonnet 5.5: no_verdict — Recounts a line that got a laugh and insults a character; no explicit judgment of the film itself. Gemini 4 Argon: positive — Praised a line as funny and cathartic; disliking the character is not a criticism of the film. Grok 4.7: positive — Praises a cathartic, laugh-out-loud line; character dislike is not film criticism. GPT-6 Luna (cheap): positive — They praise the line as a cathartic, crowd-pleasing moment, indicating enjoyment of the film. Claude Haiku 5.5 (cheap): no_verdict — VADER 0.382 Gemini 3.5 Flash-Lite (cheap): positive — Praising a great comedic moment and catharsis indicates enjoyment of the film. | + | + | − | + | 0 | 0 | + | + | + | 0 | + |
| thread #1 w 0.707 “Have you ever told a lie” “…no I don’t think I have” “Liz, come on” VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): neutral — P(op)=0.33 P(pos)=0.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.41 GPT-6.1 Sol: no_verdict — Quotes dialogue from the film without expressing an opinion about its quality. Claude Opus 5.5: no_verdict — Comment only quotes dialogue from the film without expressing any judgement of it. Claude Sonnet 5.5: no_verdict — The comment only quotes a line of dialogue and gives no judgement of the film. Gemini 4 Argon: no_verdict — The comment is simply quoting dialogue from the movie without expressing any opinion or verdict. Grok 4.7: no_verdict — Just quotes film dialogue with no judgement of the movie. GPT-6 Luna (cheap): no_verdict — Quotes dialogue without expressing an opinion about the film. Claude Haiku 5.5 (cheap): no_verdict — Just quotes dialogue from the film with no judgement of the movie itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment only quotes a line of dialogue from the movie. | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #2 w 0.577 This and the “whos on first” scene were my favorite parts VADER (old v1): positive — VADER 0.459 siebert+opinion (LIVE): positive — P(op)=0.89 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.48 GPT-6.1 Sol: positive — Calling scenes favorite parts expresses enjoyment of the film. Claude Opus 5.5: positive — Commenter names favorite parts of the film, implying enjoyment, though only lightly. Claude Sonnet 5.5: positive — Commenter calls two scenes their favorite parts, implying enjoyment of the film, though no explicit overall recommendation. Gemini 4 Argon: positive — The commenter highlights their favorite scenes, expressing enjoyment of the film without offering any criticism. Grok 4.7: positive — Calls specific scenes their favorite parts, implying enjoyment. GPT-6 Luna (cheap): positive — Calling scenes their favorites expresses enjoyment of the film. Claude Haiku 5.5 (cheap): positive — Calls two scenes their favorite parts, implying enjoyment of the film. Gemini 3.5 Flash-Lite (cheap): positive — The commenter highlights specific scenes as their favorite parts, indicating they liked the film. | + | + | 0 | + | + | + | + | + | + | + | + |
| thread #3 w 0.5 It’s a club….. that has strippers. VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): neutral — P(op)=0.30 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.50 GPT-6.1 Sol: no_verdict — Describes a club without expressing any judgment of the film. Claude Opus 5.5: no_verdict — Joke or plot observation about a setting; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Merely a factual or joking description of a setting (a strip club); expresses no opinion on the film's quality. Gemini 4 Argon: no_verdict — The comment appears to be a quote or simple observation without any opinion on the film. Grok 4.7: no_verdict — Plot description only; no judgment of the film. GPT-6 Luna (cheap): no_verdict — A brief description of the club, with no opinion about the film. Claude Haiku 5.5 (cheap): no_verdict — Joking, descriptive remark about the film's setting; no judgement of whether the film is good or recommended. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment is a simple plot description or joke without a judgment of the film. | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #4 w 0.447 Oh my god. I forgot about that. One of those things that I wished they explored more. Who were those shady ass people that he gave the rights to? Also I wished we found out more about her ex-husband. VADER (old v1): negative — VADER -0.34 siebert+opinion (LIVE): negative — P(op)=0.50 P(pos)=0.01 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.60 GPT-6.1 Sol: no_verdict — Discusses unresolved plot details and wishes for more backstory without expressing an overall judgment of the film. Claude Opus 5.5: no_verdict — Plot discussion and curiosity about unexplored elements; wishing for more isn't a clear judgement of the film. Claude Sonnet 5.5: no_verdict — Discusses plot points and wishes for more exploration of unresolved threads, but gives no clear overall judgement of the film. Gemini 4 Argon: no_verdict — Discusses unexplored plot points and unanswered questions without expressing an overall opinion on the film's quality. Grok 4.7: no_verdict — Plot curiosity about unexplored threads, no overall film judgement. GPT-6 Luna (cheap): negative — They wished the film had explored important characters and story details more, expressing dissatisfaction with what it left unexplained. Claude Haiku 5.5 (cheap): no_verdict — Commenter asks plot questions and wishes more was explored; no clear judgement of the film's quality or recommendation. Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter discusses plot points and wishes for more exploration, but expresses no overall judgment on the film. | − | − | 0 | 0 | 0 | 0 | 0 | 0 | − | 0 | 0 |
| thread #5 w 0.408 Holmes's father revealing he worked for Enron was up there for me VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): neutral — P(op)=0.19 P(pos)=0.99 Local NLI stance (deberta-v3-large): mixed — NLI p=0.36 GPT-6.1 Sol: positive — Calls the Enron revelation a standout moment, implicitly expressing appreciation for the film. Claude Opus 5.5: positive — Names a scene as one of their highlights, implying enjoyment of the film. Claude Sonnet 5.5: no_verdict — Mentions a plot detail ('up there' implies a memorable moment) but gives no clear overall judgement of the film. Gemini 4 Argon: no_verdict — The comment merely mentions a memorable detail from the content without giving an overall opinion on the film. Grok 4.7: no_verdict — Notes a memorable plot reveal without judging the film. GPT-6 Luna (cheap): positive — They single out the reveal as a standout moment, implying they enjoyed it. Claude Haiku 5.5 (cheap): no_verdict — Comment points out a plot detail about a character's backstory without expressing any judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter is discussing a specific plot point/reveal without expressing an overall judgment of the film. | 0 | 0 | 0 | + | + | 0 | 0 | 0 | + | 0 | 0 |
| thread #6 w 0.378 What actually is tough is that a lot of Enron was honestly a functioning, normal business. My friends dad worked for Enron as an engineer for a few years. No idea at all about her dad, but do kinda empathize with him specifically. Her parents were pretty clearly hurting a lot. VADER (old v1): positive — VADER 0.758 siebert+opinion (LIVE): neutral — P(op)=0.09 P(pos)=0.01 Local NLI stance (deberta-v3-large): mixed — NLI p=0.51 GPT-6.1 Sol: no_verdict — Discusses Enron employees and their families’ hardships without judging the film. Claude Opus 5.5: no_verdict — Discusses Enron and a friend's dad's experience; no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Discusses Enron background and empathy for a character's parents; offers no judgment of the film's quality or a recommendation. Gemini 4 Argon: no_verdict — Discusses real-life context and empathizes with the subjects/characters without expressing an opinion on the film itself. Grok 4.7: no_verdict — Empathizes with characters and Enron context; no film judgement. GPT-6 Luna (cheap): no_verdict — Discusses Enron and empathy for the people involved, without judging the film. Claude Haiku 5.5 (cheap): no_verdict — VADER 0.758 Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter discusses a personal connection to the subject matter without evaluating the film itself. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #7 w 0.354 There was so much pain there and neither of her parents seemed happy with Billy and her, spending the last few days on that idiocy of trying to build a Theranos 2.0 VADER (old v1): neutral — VADER 0.031 siebert+opinion (LIVE): negative — P(op)=0.98 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.37 GPT-6.1 Sol: no_verdict — Discusses characters’ pain, family relationships, and misguided actions without judging the film itself. Claude Opus 5.5: no_verdict — Discusses characters' pain and plot elements without judging the film's quality. Claude Sonnet 5.5: no_verdict — Discusses characters and plot (parents, Billy, Theranos-like venture) without judging the film's quality or recommending it; criticism targets characters, not the film. Gemini 4 Argon: no_verdict — The comment only discusses the plot and characters without giving a verdict on the film itself. Grok 4.7: no_verdict — Plot recap calling characters' scheme idiotic, no film judgement. GPT-6 Luna (cheap): no_verdict — Criticizes the characters’ choices and family situation, but gives no judgment of the film itself. Claude Haiku 5.5 (cheap): no_verdict — Describes plot and characters' emotional situations without stating whether the commenter liked or recommends the film. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses plot elements and characters without expressing a judgement on the film itself. | 0 | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #8 w 0.333 The one interaction they had made it seem pretty clear that her dad really did not like Billy VADER (old v1): positive — VADER 0.531 siebert+opinion (LIVE): negative — P(op)=0.85 P(pos)=0.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.36 GPT-6.1 Sol: no_verdict — Discusses the characters’ relationship without evaluating the film. Claude Opus 5.5: no_verdict — Comment discusses a plot detail about the dad disliking Billy; no judgement of the film. Claude Sonnet 5.5: no_verdict — Comment only discusses a plot detail about a character's relationship, with no judgment of the film itself. Gemini 4 Argon: no_verdict — The comment merely discusses character relationships and plot details without expressing an opinion on the film itself. Grok 4.7: no_verdict — Plot observation about a character interaction, no film judgement. GPT-6 Luna (cheap): no_verdict — This observes the characters’ relationship without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Plot observation about a character's father's attitude; no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses a plot point and character interaction without judging the film itself. | + | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #9 w 0.316 And maybe it was just the added discomfort of being on camera, but the dad seemed like he wanted to keep talking but was worried about irritating Billy and obviously it’s such a complex situation for the parents like the insanity of their daughters life for the past 15 years or… VADER (old v1): negative — VADER -0.929 siebert+opinion (LIVE): neutral — P(op)=0.18 P(pos)=0.01 Local NLI stance (deberta-v3-large): mixed — NLI p=0.53 GPT-6.1 Sol: no_verdict — Discusses the subjects’ family dynamics and circumstances without judging the film itself. Claude Opus 5.5: no_verdict — Discusses the parents' situation and scene dynamics without judging the film's quality. Claude Sonnet 5.5: no_verdict — Commenter discusses the parents' situation and the dad's behavior on camera, with no judgement of the film itself. Gemini 4 Argon: no_verdict — The comment analyzes the situation and subjects shown, offering no opinion on the quality of the film itself. Grok 4.7: no_verdict — Describes a tense family scene and characters without judging the film. GPT-6 Luna (cheap): no_verdict — Discusses the characters’ situation and the awkwardness of being filmed without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Describes scenes and characters' situations in the documentary without explicitly judging whether the film is good or recommended. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses plot points and character situations without expressing a judgment on the film itself. | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #10 w 0.302 The ‘Nathan for You’ fraudulent business connection not landing was simply too good VADER (old v1): negative — VADER -0.077 siebert+opinion (LIVE): negative — P(op)=0.99 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.61 GPT-6.1 Sol: positive — Calling the failed connection “simply too good” praises the film’s comedic moment. Claude Opus 5.5: positive — Praises a specific joke in the film as 'simply too good', signalling enjoyment. Claude Sonnet 5.5: no_verdict — Ambiguous remark about a 'Nathan for You' connection not landing; it reads as a joke or plot reference rather than a clear verdict on the film. Gemini 4 Argon: positive — The commenter strongly praises a specific joke or moment in the film, calling it 'simply too good'. Grok 4.7: positive — Calls the failed Nathan for You connection simply too good, praising the film. GPT-6 Luna (cheap): positive — They praise the fraudulent business connection gag as especially good. Claude Haiku 5.5 (cheap): positive — Calls a fraudulent business connection/scene 'simply too good', an approving remark about the film's content. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses a specific plot point or joke without evaluating the film itself. | − | − | 0 | + | + | 0 | + | + | + | + | 0 |
| thread #11 w 0.289 Cathartic until it became clear he was 100 times worse than initially realized VADER (old v1): negative — VADER -0.128 siebert+opinion (LIVE): negative — P(op)=0.77 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.38 GPT-6.1 Sol: no_verdict — Describes catharsis and a character's worsening moral revelation, without clearly judging the film itself. Claude Opus 5.5: no_verdict — Describes reaction to a character's revealed awfulness, not a judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Describes a character's arc or plot twist about a villain being worse than expected; no explicit judgment of the film's quality. Gemini 4 Argon: no_verdict — Describes an emotional reaction to a character's actions and plot developments without judging the quality of the film itself. Grok 4.7: negative — Initial catharsis soured once the character reveal undercut it. GPT-6 Luna (cheap): positive — Calls the experience cathartic; realizing the character was worse is not criticism of the film. Claude Haiku 5.5 (cheap): mixed — Calls it cathartic at first, then the reveal of the character being worse is a neutral-to-ambiguous reaction; no clear overall lean on the film. Gemini 3.5 Flash-Lite (cheap): mixed — Cathartic start but reveals disturbing actions, leading to a conflicted overall impression. | − | − | − | 0 | 0 | 0 | 0 | − | + | 0 | 0 |
| thread #12 w 0.277 The Charlie Kirk bit was BANANAS VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): negative — P(op)=0.96 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.77 GPT-6.1 Sol: no_verdict — Calls a specific bit wild without clearly praising or criticizing the film. Claude Opus 5.5: no_verdict — Calls one scene wild ('bananas') but gives no clear judgement of the film's quality overall. Claude Sonnet 5.5: no_verdict — Calls a scene 'bananas' but it's ambiguous whether praise or criticism; no clear overall verdict on the film. Gemini 4 Argon: no_verdict — Reacts to a specific moment as crazy ('BANANAS') without giving a clear overall judgment on the film. Grok 4.7: positive — Calls a bit bananas, enthusiastic praise of the film. GPT-6 Luna (cheap): no_verdict — Comments on one bit without expressing an overall opinion of the film. Claude Haiku 5.5 (cheap): no_verdict — Comment reacts to a specific bit's content with 'BANANAS' but gives no clear judgement of the film's quality or whether they'd recommend it. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment mentions a specific scene without expressing a clear positive or negative judgment on the film. | 0 | − | − | 0 | 0 | 0 | 0 | + | 0 | 0 | 0 |
| thread #13 w 0.267 That whole scene kept shocking me. Everything out of his mouth got more and more insane VADER (old v1): negative — VADER -0.715 siebert+opinion (LIVE): negative — P(op)=0.78 P(pos)=0.01 Local NLI stance (deberta-v3-large): negative — NLI p=0.47 GPT-6.1 Sol: no_verdict — Describes shock at a character's escalating dialogue without judging the film's quality. Claude Opus 5.5: no_verdict — Describes being shocked by a character's escalating dialogue without praising or criticising the film itself. Claude Sonnet 5.5: no_verdict — Describes a scene's shocking, insane dialogue without clearly judging the film itself; no explicit like, dislike, or recommendation. Gemini 4 Argon: no_verdict — Describes a shocking scene and character dialogue without expressing an overall opinion on the film's quality. Grok 4.7: no_verdict — Scene reaction only; no judgement of the film. GPT-6 Luna (cheap): no_verdict — Describes the scene as increasingly shocking and insane, but gives no clear opinion on the film itself. Claude Haiku 5.5 (cheap): positive — Describes a scene as shocking and the dialogue as escalating insanity, a reaction that reads as engaged appreciation rather than criticism of the film. Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter describes a shocking scene without expressing a clear verdict on the film itself. | − | − | − | 0 | 0 | 0 | 0 | 0 | 0 | + | 0 |
| thread #14 w 0.258 For real, like the fact that in that scene he straight up insinuated that he has the means and desire to physically harm all the people he wanted revenge on, only to awkwardly pause for 30 seconds before backpedaling because he probably realized that he just made a public death… VADER (old v1): negative — VADER -0.891 siebert+opinion (LIVE): negative — P(op)=0.53 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.61 GPT-6.1 Sol: no_verdict — Discusses a character’s threatening statements and behavior without judging the film’s quality. Claude Opus 5.5: no_verdict — Describes a character's alarming behavior in a scene without judging the film's quality. Claude Sonnet 5.5: no_verdict — Describes a scene and a character's alarming behavior without judging the film itself as good or bad. Gemini 4 Argon: no_verdict — The comment only describes and reacts to a character's actions in a scene without evaluating the film itself. Grok 4.7: no_verdict — Describes a character's wild dialogue without judging the film. GPT-6 Luna (cheap): no_verdict — Describes a character’s threatening behavior in a scene without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Describes a character's awkward threatening moment and amusing dialogue; no judgement of the film itself expressed. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses a specific scene and plot details without expressing a judgment or recommendation about the film. | − | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #15 w 0.25 Getting Amanda Seyfried in a room with him with a few glasses of wine was a brilliant decision. There's no way that conversation would have escalated the way it did if it was just him and Nathan. And Nathan sitting in the middle of it was just great visual comedy. As elusive… VADER (old v1): positive — VADER 0.83 siebert+opinion (LIVE): positive — P(op)=1.00 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.38 GPT-6.1 Sol: positive — Praises the dinner scene’s staging and visual comedy; contempt targets Evans, not the film. Claude Opus 5.5: positive — Praises a brilliant decision and great visual comedy; the insults target the subject, not the film. Claude Sonnet 5.5: positive — Praises the dinner scene as a brilliant decision and great visual comedy; insults aimed at Evans's character, not the film. Gemini 4 Argon: positive — Praises a 'brilliant decision' and 'great visual comedy', while expressing disgust at the documentary's subject rather than the film. Grok 4.7: positive — Praises the casting choice, visual comedy, and dinner scene as brilliant. GPT-6 Luna (cheap): positive — Calls the casting decision brilliant and the scene’s visual comedy great. Claude Haiku 5.5 (cheap): positive — VADER 0.83 Gemini 3.5 Flash-Lite (cheap): positive — Praises the acting, casting decisions, and scenes as brilliant and great. | + | + | + | + | + | + | + | + | + | + | + |
| thread #16 w 0.243 Pretty sure he’s coked out of his mind for a lot of these scenes VADER (old v1): positive — VADER 0.67 siebert+opinion (LIVE): negative — P(op)=0.94 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.70 GPT-6.1 Sol: no_verdict — Speculates about someone's drug use during scenes without judging the film's quality. Claude Opus 5.5: no_verdict — Observation about a character's or actor's state; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Observation about a character's apparent drug use; no judgement of the film's quality or recommendation. Gemini 4 Argon: no_verdict — The comment makes an observation about a person in the film without expressing an opinion on the movie itself. Grok 4.7: no_verdict — Speculates about the actor's state, not the film's quality. GPT-6 Luna (cheap): no_verdict — Speculates about someone’s state in the scenes without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Speculates about an actor's state during scenes; no judgement of the film's quality or whether they'd recommend it. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses character behavior without judging the film itself. | + | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #17 w 0.236 Glad I’m not the only one noticing the sniffling during that office scene. VADER (old v1): positive — VADER 0.459 siebert+opinion (LIVE): negative — P(op)=0.79 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.37 GPT-6.1 Sol: no_verdict — Notes sniffling during a scene without expressing an opinion on the film. Claude Opus 5.5: no_verdict — Comments on a sniffling detail in one scene without judging the film's quality. Claude Sonnet 5.5: no_verdict — Comment only notes sniffling during a scene; it gives no judgement of the film's quality or a recommendation. Gemini 4 Argon: no_verdict — The comment merely points out a small detail in a scene without expressing an opinion on the film. Grok 4.7: no_verdict — Notes a shared reaction to one scene, no film judgement. GPT-6 Luna (cheap): no_verdict — They note a detail about the office scene without evaluating the film. Claude Haiku 5.5 (cheap): no_verdict — Comments on audience sniffling during a scene, with no explicit judgement of the film's quality or recommendation. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses a minor detail in a scene without expressing a judgment on the film. | + | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #18 w 0.229 Amanda was so fucking cool in that scene. She reacted exactly how any of my friends would. VADER (old v1): positive — VADER 0.716 siebert+opinion (LIVE): positive — P(op)=0.55 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.60 GPT-6.1 Sol: positive — Praises Amanda’s scene as cool and her reaction as relatable. Claude Opus 5.5: positive — Praises a character's scene and finds her reactions realistic, implying appreciation of the film. Claude Sonnet 5.5: positive — Enthusiastically praises a character and scene as cool and relatable, implying enjoyment of the film, though no explicit overall recommendation. Gemini 4 Argon: positive — Praises a character and the realistic writing of a specific scene. Grok 4.7: positive — Praises a character’s cool, realistic reaction in a scene. GPT-6 Luna (cheap): no_verdict — Praises Amanda’s behavior in one scene, but gives no overall judgment of the film. Claude Haiku 5.5 (cheap): positive — Praises a character's scene as cool and relatable; enthusiastic tone implies approval of the film's material. Gemini 3.5 Flash-Lite (cheap): no_verdict — Discusses a character's action in a specific scene without expressing an overall opinion on the film. | + | + | 0 | + | + | + | + | + | 0 | + | 0 |
| thread #19 w 0.224 She was so cool throughout. It was extremely cathartic to see her be as openly baffled and disgusted by Billy as I was. On the other hand, I'm not sure how to express this but it was interesting to see her reaction to Elizabeth blaming the show for making her life harder in… VADER (old v1): positive — VADER 0.634 siebert+opinion (LIVE): positive — P(op)=0.56 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.88 GPT-6.1 Sol: positive — Praises Amanda as a major highlight and values her cathartic presence; discomfort concerns ethical complexity, not filmmaking. Claude Opus 5.5: positive — Praises Amanda as a major highlight and the film as cathartic; its complexity is framed as interesting, not a flaw. Claude Sonnet 5.5: no_verdict — Mostly analyzes Amanda's reactions and the film's events; praise is for a person in the doc, with no clear overall verdict on the film. Gemini 4 Argon: positive — The commenter finds the documentary's subject matter cathartic, interesting, and nuanced, praising one interviewee as a major highlight. Grok 4.7: positive — Calls Amanda a major highlight and is glad she was there; complexity is not criticism. GPT-6 Luna (cheap): positive — Calls Amanda a major highlight of the documentary and says they are extremely glad she was included. Claude Haiku 5.5 (cheap): positive — VADER 0.634 Gemini 3.5 Flash-Lite (cheap): positive — Praises Amanda as a major highlight and expresses gladness she was in the movie despite the complex feelings. | + | + | 0 | + | + | 0 | + | + | + | + | + |
| thread #20 w 0.218 Billy: I have anger in my heart and I’ll be seeking vengeance on everyone Nathan: Everyone?! VADER (old v1): positive — VADER 0.2 siebert+opinion (LIVE): neutral — P(op)=0.08 P(pos)=0.01 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.50 GPT-6.1 Sol: no_verdict — Quotes character dialogue without expressing an opinion on the film. Claude Opus 5.5: no_verdict — Quotes dialogue from the film without expressing any judgement of its quality. Claude Sonnet 5.5: no_verdict — The comment is just a quoted exchange of dialogue from the film, with no judgement of the film itself. Gemini 4 Argon: no_verdict — The comment is simply a quote or dialogue joke with no judgment on the film. Grok 4.7: no_verdict — Quoted dialogue only; no judgment of the film. GPT-6 Luna (cheap): no_verdict — A quoted exchange about vengeance, with no opinion about the film itself. Claude Haiku 5.5 (cheap): no_verdict — Only quotes a line of dialogue with no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment only quotes lines of dialogue from the movie. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #21 w 0.213 I've never felt so uncertain about my feelings after finishing a movie. I truly have no idea where the act ends and reality begins VADER (old v1): negative — VADER -0.288 siebert+opinion (LIVE): neutral — P(op)=0.37 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.56 GPT-6.1 Sol: no_verdict — Expresses uncertainty and blurred boundaries between acting and reality, without judging the film positively or negatively. Claude Opus 5.5: no_verdict — Expresses uncertainty and confusion about the film's ambiguity without judging its quality. Claude Sonnet 5.5: no_verdict — Commenter expresses uncertainty and confusion about their feelings and the film's reality; no clear like or dislike verdict. Gemini 4 Argon: no_verdict — The commenter explicitly states they are uncertain about their feelings and gives no positive or negative judgment. Grok 4.7: no_verdict — Uncertain feelings only; no clear like or dislike of the film. GPT-6 Luna (cheap): no_verdict — They describe feeling uncertain about the boundary between acting and reality, but give no clear evaluation of the film. Claude Haiku 5.5 (cheap): no_verdict — Describes an uncertain emotional reaction and confusion about reality, but gives no clear like or dislike of the film. Gemini 3.5 Flash-Lite (cheap): no_verdict — Expresses uncertainty about the film's impact without stating a clear like or dislike. | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #22 w 0.209 Oh so the Nathan Fielder movie is a Nathan Fielder movie then. VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): neutral — P(op)=0.04 P(pos)=0.01 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.79 GPT-6.1 Sol: no_verdict — Comments on the film matching Nathan Fielder’s style without expressing approval or disapproval. Claude Opus 5.5: no_verdict — Wry observation that the film matches Fielder's style; no clear judgement of quality. Claude Sonnet 5.5: no_verdict — Wry remark that the film is typical of Nathan Fielder's style; no clear judgement of quality or recommendation. Gemini 4 Argon: no_verdict — The comment makes a neutral observation about the creator's style without expressing a positive or negative opinion. Grok 4.7: no_verdict — Wry joke about the director's style, no film judgement. GPT-6 Luna (cheap): no_verdict — It identifies the film as characteristic of Nathan Fielder without expressing whether the commenter liked it. Claude Haiku 5.5 (cheap): no_verdict — Quips that the film is typical of Nathan Fielder's style; no clear judgement of whether the commenter liked it. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment is a stylistic observation about the director's style without expressing a clear judgment on the film. | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #23 w 0.204 It’s a Nathan Fielder formula without the Nathan Fielder character. At least, that was my takeaway. VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): neutral — P(op)=0.16 P(pos)=0.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.53 GPT-6.1 Sol: no_verdict — Compares the film’s formula to Nathan Fielder’s work without expressing approval or disappointment. Claude Opus 5.5: no_verdict — Compares the film's formula to Nathan Fielder's work without clearly judging its quality. Claude Sonnet 5.5: no_verdict — Describes the film as a Nathan Fielder formula without his character; ambiguous observation, no clear judgement of quality or recommendation. Gemini 4 Argon: no_verdict — The commenter compares the film's style to Nathan Fielder without expressing whether they liked or disliked it. Grok 4.7: no_verdict — Observational comparison only; no clear like or dislike of the film. GPT-6 Luna (cheap): negative — Calling it a Nathan Fielder formula without the defining character suggests the film feels lacking. Claude Haiku 5.5 (cheap): negative — Commenter's takeaway frames the film as a weaker imitation of Fielder's formula, implying disappointment with it. Gemini 3.5 Flash-Lite (cheap): no_verdict — Describes the film's style and comparison without expressing a clear like or dislike. | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | − | − | 0 |
| thread #24 w 0.2 Yeah, I loved watching him continue to press follow up questions and pulling on each thread. Especially when he’s asking Billy about the first time they met and fell in love. “What thought did you just have?” VADER (old v1): positive — VADER 0.883 siebert+opinion (LIVE): neutral — P(op)=0.06 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.90 GPT-6.1 Sol: positive — Expresses enjoyment of watching the character persistently ask probing follow-up questions. Claude Opus 5.5: positive — Commenter says they loved watching the character's questioning scenes, expressing enjoyment of the film. Claude Sonnet 5.5: positive — Commenter says they loved watching the character's follow-up questioning and singles out a specific scene, which expresses enjoyment of the film. Gemini 4 Argon: positive — The commenter explicitly states they loved watching the interviewing/questioning aspect of the film. Grok 4.7: positive — Loved the questioning scene and how threads were pulled. GPT-6 Luna (cheap): positive — They explicitly loved watching the film’s questioning and relationship scene. Claude Haiku 5.5 (cheap): positive — Commenter says they loved watching the interviewer's follow-up questioning and specific scene moments, expressing enjoyment of the film. Gemini 3.5 Flash-Lite (cheap): positive — The commenter explicitly states they loved watching the scene unfold. | + | 0 | 0 | + | + | + | + | + | + | + | + |
| thread #25 w 0.196 I loved how he did not let up. VADER (old v1): positive — VADER 0.599 siebert+opinion (LIVE): positive — P(op)=0.93 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.66 GPT-6.1 Sol: positive — Expresses love for the film's relentless approach. Claude Opus 5.5: positive — Commenter says they loved the relentlessness, which is praise for the film's approach. Claude Sonnet 5.5: positive — Commenter explicitly says they loved how relentless he was, which is praise for the film's intensity. Gemini 4 Argon: positive — The commenter explicitly states they 'loved' an aspect of the performance/character in the film. Grok 4.7: positive — Explicitly loved an aspect of the film's intensity. GPT-6 Luna (cheap): positive — They explicitly loved an aspect of the film. Claude Haiku 5.5 (cheap): positive — Commenter says they loved how the film did not let up, clear praise. Gemini 3.5 Flash-Lite (cheap): positive — The commenter explicitly states they loved the relentless nature of the film. | + | + | 0 | + | + | + | + | + | + | + | + |
| thread #26 w 0.192 It’s actually the unofficial 3rd season of The Rehearsal guest starring Amanda Seyfried VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): neutral — P(op)=0.09 P(pos)=0.99 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.70 GPT-6.1 Sol: no_verdict — Joking comparison to The Rehearsal without expressing whether the film is good or bad. Claude Opus 5.5: no_verdict — Joke comparing the film to The Rehearsal; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — A joking comparison to The Rehearsal with a cast mention; it gives no clear judgement of the film's quality. Gemini 4 Argon: no_verdict — The comment makes a joke comparing the premise to a TV show without offering a verdict on the film. Grok 4.7: no_verdict — Joke comparing the film to another show, no judgement. GPT-6 Luna (cheap): no_verdict — A joking comparison to another show, with no opinion about the film’s quality. Claude Haiku 5.5 (cheap): no_verdict — Joking comparison to a TV show; no clear judgement of the film's quality or whether to watch it. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment is a joke comparing the film's style to another show, not judging the movie itself. | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #27 w 0.189 Before seeing it, I think I just assumed it was going to end with Elizabeth going to prison and got surprised when it happened probably halfway through. Turns out there was so much more to this than I could have imagined. VADER (old v1): negative — VADER -0.34 siebert+opinion (LIVE): neutral — P(op)=0.21 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.88 GPT-6.1 Sol: positive — Expresses appreciative surprise at the film's unexpected scope and depth beyond their initial expectations. Claude Opus 5.5: positive — Pleasantly surprised the film offered far more than expected; implies appreciation. Claude Sonnet 5.5: positive — Commenter was pleasantly surprised the film had far more depth than expected, implying a favorable view, though no explicit recommendation. Gemini 4 Argon: positive — The commenter expresses pleasant surprise that the film had much more depth and story than they initially expected. Grok 4.7: positive — Surprised by depth; film exceeded expectations with more than imagined. GPT-6 Luna (cheap): positive — They express pleasant surprise that the film offered much more than they expected. Claude Haiku 5.5 (cheap): positive — Commenter was surprised and says there was 'so much more to this than I could have imagined', a favorable reaction to the film. Gemini 3.5 Flash-Lite (cheap): positive — The commenter expresses pleasant surprise and admiration for the film's unexpected depth and scope. | − | 0 | 0 | + | + | + | + | + | + | + | + |
| thread #28 w 0.186 The movie’s slow reveal that Billy is just as much of, if not MORE of a manipulative psychopath than Elizabeth is was fucking phenomenal. At first I was just bought in that he was a dude she was scamming that was too caught up thinking with his dick to see it. But good god he… VADER (old v1): positive — VADER 0.223 siebert+opinion (LIVE): positive — P(op)=1.00 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.97 GPT-6.1 Sol: positive — Calls the reveal phenomenal and the film incredible art, praising its ambiguity, cinematography, and soundtrack. Claude Opus 5.5: positive — Calls the slow reveal phenomenal, the film an incredible work of art, and praises the cinematography and soundtrack. Claude Sonnet 5.5: positive — Calls the slow reveal phenomenal and the film an incredible work of art, praising cinematography and soundtrack. Gemini 4 Argon: positive — The commenter explicitly praises the film as 'fucking phenomenal,' 'an incredible work of art,' and 'so goddamn good.' Grok 4.7: positive — Calls the film phenomenal and an incredible work of art, praising craft and soundtrack. GPT-6 Luna (cheap): positive — Calls the film phenomenal and an incredible work of art, praising its suspense, ambiguity, cinematography, and soundtrack. Claude Haiku 5.5 (cheap): positive — VADER 0.223 Gemini 3.5 Flash-Lite (cheap): positive — The commenter calls the movie 'fucking phenomenal', an 'incredible work of art', and praises its cinematography and soundtrack. | + | + | + | + | + | + | + | + | + | + | + |
| thread #29 w 0.183 I think that, to an extent, their relationship is an act. The split in the financial stuff seems intentional so that Holmes doesn't have to pay back the people she defrauded. For the other stuff, sure maybe she's a total victim but I think it falls pretty clearly into a pattern… VADER (old v1): positive — VADER 0.927 siebert+opinion (LIVE): neutral — P(op)=0.32 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.44 GPT-6.1 Sol: no_verdict — Discusses Holmes and Billy’s relationship and alleged deception without evaluating the film itself. Claude Opus 5.5: no_verdict — Analyzes Holmes and Billy's relationship and motives; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Discusses the subjects' relationship and theories about Holmes and Billy; offers no judgement of the film itself. Gemini 4 Argon: no_verdict — Discusses the real-life subjects and their relationship dynamics without expressing an opinion on the film itself. Grok 4.7: no_verdict — Character and plot theory only; no judgement of the film. GPT-6 Luna (cheap): no_verdict — They discuss the characters’ relationship and motives without judging the film itself. Claude Haiku 5.5 (cheap): no_verdict — VADER 0.927 Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses plot theories and character dynamics without expressing a verdict on the film itself. | + | 0 | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #30 w 0.18 When her dad just confidently and casually said where it all went wrong… that was shocking. And it really made me feel like genuine damage was done to her brain and that she never recovered, and now doesn’t have a single original thought in her head outside of what her partner… VADER (old v1): positive — VADER 0.402 siebert+opinion (LIVE): negative — P(op)=0.73 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.50 GPT-6.1 Sol: no_verdict — Discusses a shocking scene and its implications for a character without judging the film. Claude Opus 5.5: no_verdict — Discusses a shocking scene and character's damage; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Discusses a plot moment and character impact without explicitly judging the film's quality or recommending it. Gemini 4 Argon: no_verdict — Discusses a character and plot reaction without offering a judgement on the quality of the film itself. Grok 4.7: no_verdict — Describes a shocking plot moment and character damage, no film judgement. GPT-6 Luna (cheap): no_verdict — Describes a shocking scene and its effect on a character, but gives no judgment of the film itself. Claude Haiku 5.5 (cheap): no_verdict — VADER 0.402 Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter discusses plot points and character impact without expressing a clear judgment on the film itself. | + | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #31 w 0.177 I desperately wish we'd gotten to hear more from him. I could really see how many of her behaviors could have stemmed from that - the relationship patterns, selective forgetfulness, and weaponized helplessness VADER (old v1): negative — VADER -0.459 siebert+opinion (LIVE): neutral — P(op)=0.15 P(pos)=0.04 Local NLI stance (deberta-v3-large): mixed — NLI p=0.59 GPT-6.1 Sol: no_verdict — Discusses character behavior and wishes for more character perspective without judging the film overall. Claude Opus 5.5: no_verdict — Character analysis and a wish for more of one character; no clear judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Discusses character analysis and wishes for more of a character, but gives no clear overall judgment of the film. Gemini 4 Argon: no_verdict — The comment analyzes character psychology and wishes for more focus on one character, without judging the film's overall quality. Grok 4.7: no_verdict — Character analysis and wanting more backstory, no film judgement. GPT-6 Luna (cheap): negative — They express disappointment that the film did not let him speak more, criticizing its presentation of his perspective. Claude Haiku 5.5 (cheap): no_verdict — Discusses character motivations and wishes for more story content, but gives no clear judgement of the film's quality or recommendation. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses character behavior and plot psychology without expressing any judgment or verdict on the film itself. | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | − | 0 | 0 |
| thread #32 w 0.174 Yeah it's actually criminal that Nathan never got the parents' full responses. But Billy clearly was preventing it. The question is why- what were they gonna say? VADER (old v1): positive — VADER 0.421 siebert+opinion (LIVE): negative — P(op)=0.58 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.45 GPT-6.1 Sol: no_verdict — Discusses characters’ actions and speculates about withheld responses without judging the film itself. Claude Opus 5.5: no_verdict — Plot discussion and speculation about characters; 'criminal' refers to a story event, not a judgement of the film. Claude Sonnet 5.5: no_verdict — Plot discussion and speculation about characters' motives and unanswered questions, with no judgement of the film's quality. Gemini 4 Argon: no_verdict — The comment only discusses plot points and character actions without giving a verdict on the film itself. Grok 4.7: no_verdict — Plot speculation about characters, no film judgement. GPT-6 Luna (cheap): no_verdict — Discusses a plot question without expressing an opinion about the film. Claude Haiku 5.5 (cheap): no_verdict — Discusses plot details and asks a question about characters; no judgement of the film's quality is expressed. Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter discusses plot details and asks a question without expressing an opinion on the film. | + | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #33 w 0.171 He can control her, but not them. Same behavior he displayed when Nathan tried to interview the business partner after the time jump. VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): neutral — P(op)=0.04 P(pos)=0.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.47 GPT-6.1 Sol: no_verdict — Discusses character behavior and plot events without evaluating the film. Claude Opus 5.5: no_verdict — Character behavior analysis linking plot events; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Plot/character analysis about a character's controlling behavior; no judgement of the film's quality or recommendation. Gemini 4 Argon: no_verdict — The comment only discusses plot and character behavior without expressing an opinion on the film's quality. Grok 4.7: no_verdict — Plot and character behavior analysis only, no film judgement. GPT-6 Luna (cheap): no_verdict — Observes a character's behavior and draws a plot comparison without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Plot discussion about a character's behavior, with no judgement of the film's quality or whether the commenter liked it. Gemini 3.5 Flash-Lite (cheap): no_verdict — Plot discussion and character behavior analysis without expressing an opinion on the film. | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #34 w 0.169 I agree, people are forgetting Liz and Billy are trying to control the narrative the whole time. They know they are being filmed. It's all a performance. When Liz says she has no idea who Nathan Fielder is. Come on. You really did zero research about this person who is going to… VADER (old v1): positive — VADER 0.392 siebert+opinion (LIVE): positive — P(op)=0.98 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.70 GPT-6.1 Sol: positive — Explicitly loved the movie’s duel between two masters of deception. Claude Opus 5.5: positive — Analyzes characters' performance, then explicitly says 'that's what I loved about this movie.' Claude Sonnet 5.5: positive — Commenter explicitly says 'that's what I loved about this movie' and praises it as a duel of two masters of deception. Gemini 4 Argon: positive — The commenter explicitly states what they 'loved about this movie' and praises the dynamic between the subjects. Grok 4.7: positive — Explicitly loved the film as a duel of deception. GPT-6 Luna (cheap): positive — The commenter explicitly says they loved the movie and praises its central deception duel. Claude Haiku 5.5 (cheap): positive — VADER 0.392 Gemini 3.5 Flash-Lite (cheap): positive — The commenter explicitly states 'that's what I loved about this movie', praising its dynamic and performances. | + | + | 0 | + | + | + | + | + | + | + | + |
| thread #35 w 0.167 The only time I believed her 100% was when they were in the prison talking and she said their souls are intertwined or whatever. I think they are one and the same - 2 giant grifters. For nearly all of the questions Nathan asks, she somehow manages to give long nonanswers that… VADER (old v1): positive — VADER 0.911 siebert+opinion (LIVE): negative — P(op)=1.00 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.78 GPT-6.1 Sol: no_verdict — Discusses the subjects’ dishonesty and motives without evaluating the film itself. Claude Opus 5.5: no_verdict — Discusses the subject's credibility and behavior; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Commenter critiques the subject's honesty and discusses her motives, not the film's quality; no clear judgment on the film itself. Gemini 4 Argon: no_verdict — The comment analyzes the documentary's subjects and their motives without expressing an opinion on the quality of the film itself. Grok 4.7: no_verdict — Discusses the subject's honesty and motives, not whether the film works. GPT-6 Luna (cheap): no_verdict — They judge the subject’s credibility and behavior, but express no opinion on the film itself. Claude Haiku 5.5 (cheap): no_verdict — VADER 0.911 Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses character motivations and plot details without evaluating the film itself. | + | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #36 w 0.164 Beyond thrilled Billy’s coke snorts were left in by the editor. Dude can get royally fucked. VADER (old v1): negative — VADER -0.361 siebert+opinion (LIVE): negative — P(op)=0.99 P(pos)=0.03 Local NLI stance (deberta-v3-large): negative — NLI p=0.50 GPT-6.1 Sol: no_verdict — Celebrates retaining Billy’s coke snorts and insults him, without judging the film overall. Claude Opus 5.5: no_verdict — Mainly hostility toward Billy and approval of one editing choice; no overall judgement of the film. Claude Sonnet 5.5: no_verdict — Comment is about an editing choice and hostility toward a person (Billy); no clear overall judgment of the film itself. Gemini 4 Argon: positive — Praises a specific editing choice in the film while expressing strong dislike for the subject. Grok 4.7: positive — Thrilled the editor kept Billy's coke snorts; character hate isn't film criticism. GPT-6 Luna (cheap): positive — They enthusiastically praise the editor for leaving in the coke snorts, expressing approval of that choice. Claude Haiku 5.5 (cheap): positive — Commenter is 'beyond thrilled' about a detail left in the film, an enthusiastic approval of the movie's content. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses a specific detail about a character rather than providing a verdict on the film. | − | − | − | 0 | 0 | 0 | + | + | + | + | 0 |
| thread #37 w 0.162 That they met in a WeWork was top tier comedy VADER (old v1): positive — VADER 0.511 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.69 GPT-6.1 Sol: positive — Praises the film’s WeWork meeting scene as top-tier comedy. Claude Opus 5.5: positive — Praises a specific comedic detail in the film as 'top tier comedy,' implying enjoyment. Claude Sonnet 5.5: positive — Praises a specific detail as 'top tier comedy', implying enjoyment of the film's humor, though no explicit overall recommendation. Gemini 4 Argon: positive — Praises a specific comedic detail in the film as 'top tier comedy'. Grok 4.7: positive — Calls a scene top tier comedy, praising the film's humor. GPT-6 Luna (cheap): positive — They praise the WeWork meeting as top-tier comedy, expressing enjoyment of the film’s humor. Claude Haiku 5.5 (cheap): no_verdict — Joke about a plot detail; no clear judgement of the film's quality or recommendation. Gemini 3.5 Flash-Lite (cheap): positive — Praising a specific aspect of the film as top tier comedy. | + | + | 0 | + | + | + | + | + | + | 0 | + |
| thread #38 w 0.16 I was the only one laughing when that was shown. Along with the mention of the gas station episode of Nathan For You. Had me giggling loud remembering it and how that was the first time, for me at least, where Nathan broke character in that show VADER (old v1): positive — VADER 0.44 siebert+opinion (LIVE): neutral — P(op)=0.03 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.87 GPT-6.1 Sol: no_verdict — Describes laughing at a reference and remembering Nathan For You, without judging the film overall. Claude Opus 5.5: no_verdict — Shares a personal reaction and Nathan For You trivia without judging the film's quality. Claude Sonnet 5.5: no_verdict — Commenter shares a personal reaction to a specific moment and a Nathan For You reference, but gives no overall judgement of the film. Gemini 4 Argon: no_verdict — The commenter shares an anecdote about laughing at specific references without giving an overall verdict on the film. Grok 4.7: no_verdict — Laughs at a scene and recalls Nathan For You; no film judgement. GPT-6 Luna (cheap): no_verdict — They describe laughing at a reference and recalling a TV episode, without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Describes laughing at a scene and recalling a Nathan For You episode; no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses a funny moment and a television show reference without expressing a verdict on the film. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #39 w 0.158 Dude is a coked up lowlife. VADER (old v1): negative — VADER -0.361 siebert+opinion (LIVE): negative — P(op)=0.91 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.66 GPT-6.1 Sol: no_verdict — Insults a character without expressing any judgment of the film itself. Claude Opus 5.5: no_verdict — Describes a character negatively; no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Describes a character negatively, but offers no judgment of the film itself. Gemini 4 Argon: no_verdict — The comment describes a character or person without offering a judgment on the quality of the film itself. Grok 4.7: no_verdict — Character insult only; no judgement of the film. GPT-6 Luna (cheap): no_verdict — Insults a character but gives no opinion on the film. Claude Haiku 5.5 (cheap): no_verdict — Describes a character as a coked-up lowlife without any judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses a character without expressing a verdict on the film itself. | − | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #40 w 0.156 The Meta of the prison scene of Amanda playing Elizabeth saying she never wants someone to play her again 👌 VADER (old v1): positive — VADER 0.721 siebert+opinion (LIVE): neutral — P(op)=0.09 P(pos)=0.97 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.60 GPT-6.1 Sol: positive — The approving emoji praises the prison scene’s meta commentary and layered performance. Claude Opus 5.5: positive — The 👌 emoji signals appreciation of the meta prison scene, a mild approving remark. Claude Sonnet 5.5: no_verdict — Comment points out a meta detail about a scene without stating whether they liked or disliked the film. Gemini 4 Argon: positive — The commenter praises a clever meta moment in a scene, expressing approval with an 'OK hand' emoji. Grok 4.7: positive — Approves the meta prison scene with an OK emoji. GPT-6 Luna (cheap): positive — The approving emoji praises the film’s meta prison-scene moment. Claude Haiku 5.5 (cheap): no_verdict — Comment discusses a specific scene's meta-level irony without expressing whether the commenter likes or recommends the film. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses a specific meta scene without expressing an overall opinion or recommendation on the film. | + | 0 | 0 | + | + | 0 | + | + | + | 0 | 0 |
| thread #41 w 0.154 I genuinely hope she gets a nomination for Best Actress. She deserves it. It would piss off Billy so fucking much. VADER (old v1): positive — VADER 0.66 siebert+opinion (LIVE): positive — P(op)=0.44 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.52 GPT-6.1 Sol: no_verdict — Supports an actress’s nomination and anticipates Billy’s reaction, without judging the film itself. Claude Opus 5.5: positive — Says the lead actress deserves a Best Actress nomination, which is strong praise for the film's performance. Claude Sonnet 5.5: positive — Commenter praises the lead actress's performance as award-worthy, implying a favorable view of the film's acting, though no overall verdict is stated. Gemini 4 Argon: positive — Praises the lead actress's performance, stating she deserves a Best Actress nomination. Grok 4.7: positive — Praises the lead performance as Oscar-worthy and deserved. GPT-6 Luna (cheap): no_verdict — Praises the actress’s performance but gives no opinion on the film itself. Claude Haiku 5.5 (cheap): positive — Commenter praises the lead actress and hopes she gets a Best Actress nomination, implying strong approval of her performance in the film. Gemini 3.5 Flash-Lite (cheap): positive — Praising the lead performance and hoping she gets an award nomination indicates a positive verdict. | + | + | 0 | 0 | + | + | + | + | 0 | + | + |
| thread #42 w 0.152 I know Billy wasn’t acting but that scene at the dinner table between them was phenomenal. He was so perfectly hate-able. VADER (old v1): positive — VADER 0.861 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.46 GPT-6.1 Sol: positive — Calls the dinner table scene phenomenal and praises Billy's perfectly hateable portrayal. Claude Opus 5.5: positive — Calls the dinner scene phenomenal; hating the character is praise of the performance, not criticism of the film. Claude Sonnet 5.5: positive — Praises the dinner table scene as phenomenal and the character as perfectly hate-able; this is praise of the film's content, not criticism. Gemini 4 Argon: positive — The commenter praises a specific scene as phenomenal and compliments how effectively hateable the character is. Grok 4.7: positive — Praises a dinner scene as phenomenal and the character as perfectly hateable. GPT-6 Luna (cheap): positive — Calls the dinner-table scene phenomenal; disliking Billy is about the character, not the film. Claude Haiku 5.5 (cheap): positive — Praises a scene as phenomenal and the performance as perfectly hate-able, a clear compliment to the film's acting. Gemini 3.5 Flash-Lite (cheap): positive — Praises the dinner table scene as phenomenal and the acting as perfect. | + | + | 0 | + | + | + | + | + | + | + | + |
| thread #43 w 0.151 This is what he's like ON CAMERA. Wtf is wrong with these people VADER (old v1): negative — VADER -0.66 siebert+opinion (LIVE): negative — P(op)=0.54 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.92 GPT-6.1 Sol: no_verdict — Reacts to people's behavior on camera without judging the film itself. Claude Opus 5.5: no_verdict — Expresses outrage at a person's on-camera behavior, not a judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Expresses shock at a person's behavior on camera; criticizes people, not the film itself, and gives no verdict on its quality. Gemini 4 Argon: no_verdict — The comment reacts to the characters/subjects in the film without expressing a judgment on the movie's quality. Grok 4.7: no_verdict — Reacts to people shown, not the film itself. GPT-6 Luna (cheap): no_verdict — Reacts to someone’s on-camera behavior without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Comment reacts to the subject's on-camera behaviour, not a judgement of the film itself; no clear verdict on the film. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses characters or real people, not the film itself. | − | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #44 w 0.149 He was high and got carried away. You can see the moment he remembers the cameras are there. VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): positive — P(op)=0.51 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.62 GPT-6.1 Sol: no_verdict — Interprets a character's behavior without evaluating the film. Claude Opus 5.5: no_verdict — Describes a character's behavior in a scene; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Describes or speculates about a moment in the film without giving any judgment of the film itself. Gemini 4 Argon: no_verdict — The comment merely describes a character's actions in a scene without expressing any opinion on the film's quality. Grok 4.7: no_verdict — Describes a character's behavior, no judgment of the film. GPT-6 Luna (cheap): no_verdict — Describes a character’s behavior without expressing an opinion about the film. Claude Haiku 5.5 (cheap): no_verdict — Describes a plot moment and character behavior without any judgement of the film's quality or whether they'd recommend it. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses a plot detail/character action without judging the film itself. | 0 | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #45 w 0.147 Nathan: you would harm people? Billy: I'd do what I need to... (Nathan shocked) ... (Amanda shocked) ... (looks at camera) ... Billy: But then I remember Liz and I calm down. Fucking psychopath VADER (old v1): negative — VADER -0.153 siebert+opinion (LIVE): neutral — P(op)=0.30 P(pos)=0.01 Local NLI stance (deberta-v3-large): negative — NLI p=0.60 GPT-6.1 Sol: no_verdict — Quotes dialogue and calls a character a psychopath without judging the film itself. Claude Opus 5.5: no_verdict — Quotes a scene and calls a character a psychopath; no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Recounts a scene and calls the character Billy a psychopath, which is character commentary, not a judgement of the film. Gemini 4 Argon: no_verdict — Quotes a scene and calls a character a psychopath without judging the quality of the film itself. Grok 4.7: no_verdict — Quotes dialogue and calls a character a psychopath, no film judgement. GPT-6 Luna (cheap): no_verdict — Quotes dialogue and calls a character a psychopath, without judging the film itself. Claude Haiku 5.5 (cheap): no_verdict — Quotes a scene and calls the character a psychopath; no judgement of the film itself is expressed. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment only quotes movie dialogue and reacts to a character without judging the film itself. | − | 0 | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #46 w 0.146 Amanda really did do a great job baiting him. Fucking psycho. VADER (old v1): positive — VADER 0.625 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=0.99 Local NLI stance (deberta-v3-large): positive — NLI p=0.42 GPT-6.1 Sol: no_verdict — Discusses Amanda’s manipulation and character without judging the film itself. Claude Opus 5.5: no_verdict — Comments on a character's behavior and calls her a psycho; no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Comments on a character's plot actions and calls someone a psycho; no explicit judgment of the film itself. Gemini 4 Argon: no_verdict — The comment only reacts to a character's actions and calls them a psycho without evaluating the film. Grok 4.7: no_verdict — Character reaction only; no judgement of the film. GPT-6 Luna (cheap): no_verdict — Comments on Amanda’s actions and character, not the film’s quality. Claude Haiku 5.5 (cheap): no_verdict — Calls a character a psycho and praises her acting-like manipulation; no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses a character's actions without expressing a judgment on the film itself. | + | + | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #47 w 0.144 Also seeing the real Amanda and how much she sort of resembles Holmes was nuts. Real Amanda also acted as a sort of proxy for us the audience, because she was wrestling with her own feelings about it the whole time, going on the journey of discovery with all of us that Nathan… VADER (old v1): positive — VADER 0.128 siebert+opinion (LIVE): positive — P(op)=0.51 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.48 GPT-6.1 Sol: positive — Praises the film’s staggering meta layers and Amanda’s role in the audience’s journey of discovery. Claude Opus 5.5: positive — Calls the film's meta achievement 'staggering' and the real Amanda's resemblance 'nuts', both admiring. Claude Sonnet 5.5: positive — Commenter praises the film's meta level as 'staggering' and admires the real Amanda's role as audience proxy, implying approval. Gemini 4 Argon: positive — Praises the film's execution, calling the level of meta it achieved 'staggering' and a detail 'nuts'. Grok 4.7: positive — Praises the film's staggering meta level and audience proxy. GPT-6 Luna (cheap): positive — Calls the film’s meta approach staggering and praises Amanda’s role in the audience’s journey. Claude Haiku 5.5 (cheap): positive — VADER 0.128 Gemini 3.5 Flash-Lite (cheap): positive — The commenter praises the film's meta elements, calling the achievement 'staggering'. | + | + | 0 | + | + | + | + | + | + | + | + |
| thread #48 w 0.143 I won’t lie either that I love how she called out more bullshit than anyone else in the movie. She’s the only person who actually decided to get under Billy’s, and Elizabeth’s, skin. VADER (old v1): neutral — VADER 0.028 siebert+opinion (LIVE): positive — P(op)=1.00 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.39 GPT-6.1 Sol: no_verdict — Praises a character’s confrontational behavior without expressing an overall judgment of the film. Claude Opus 5.5: no_verdict — Praises a character's actions within the story, but gives no judgement of the film overall. Claude Sonnet 5.5: no_verdict — Praises a character's actions within the plot, not a judgment of the film itself. Gemini 4 Argon: no_verdict — Expresses appreciation for a specific character's actions in the plot without giving an overall verdict on the film itself. Grok 4.7: positive — Praises a character for calling out others, showing enjoyment of the film. GPT-6 Luna (cheap): positive — They express love for a character’s actions in the film, signaling a positive reaction to that aspect. Claude Haiku 5.5 (cheap): no_verdict — VADER 0.028 Gemini 3.5 Flash-Lite (cheap): positive — The commenter expresses clear enjoyment of the character's actions in the film. | 0 | + | + | 0 | 0 | 0 | 0 | + | + | 0 | + |
| thread #49 w 0.141 Amanda was my hero in the movie. As much as she believed Elizabeth was where she belonged and continued to be suspicious of her, she wanted to end that controling, abusive marriage. Her reactions from seeing Billy take the bowl to bringing up Charlie Kirk - chef's kiss. VADER (old v1): negative — VADER -0.077 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.60 GPT-6.1 Sol: positive — Praises Amanda’s characterization and reactions, calling them “chef’s kiss,” conveying appreciation for the film. Claude Opus 5.5: positive — Praises Amanda's character and her reactions with 'chef's kiss', implying enjoyment of the film. Claude Sonnet 5.5: positive — Commenter praises Amanda's character and specific scenes enthusiastically ('chef's kiss', 'my hero'), implying appreciation of the film, though no explicit overall verdict. Gemini 4 Argon: positive — Praises a character and describes her reactions in the film as "chef's kiss". Grok 4.7: positive — Praises Amanda as a hero and calls her reactions chef's kiss. GPT-6 Luna (cheap): positive — They praise Amanda and describe her reactions as “chef’s kiss,” expressing appreciation for the film’s characterization. Claude Haiku 5.5 (cheap): positive — Calls the character their hero and praises her reactions with 'chef's kiss', signaling approval of the film's performance and story. Gemini 3.5 Flash-Lite (cheap): positive — Praises the character's actions and calls specific scenes 'chef's kiss'. | − | + | 0 | + | + | + | + | + | + | + | + |
| thread #50 w 0.14 I walked away thinking the real Amanda seems cool as shit. She was so down for what Nathan wanted to do and I love the conviction she had about Charlie. My theater applauded when she insulted him again with zero hesitation. VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): positive — P(op)=0.85 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.76 GPT-6.1 Sol: no_verdict — Praises Amanda’s personality and conviction, and describes audience reactions, without judging the film itself. Claude Opus 5.5: positive — Enthusiastic about a character and describes the theater applauding a scene; implies enjoyment of the film. Claude Sonnet 5.5: positive — Enjoyed the film, loved the character conviction, and noted the theater applauded; overall favorable reaction to the film. Gemini 4 Argon: no_verdict — Discusses a character/subject and the theater's reaction, without giving an overall verdict on the film itself. Grok 4.7: positive — Praises Amanda and a scene the theater applauded, showing enjoyment. GPT-6 Luna (cheap): no_verdict — Praises Amanda and an audience reaction, but gives no judgment of the film itself. Claude Haiku 5.5 (cheap): positive — VADER 0.0 Gemini 3.5 Flash-Lite (cheap): positive — Praises the characters and notes the theater applauded, indicating a favorable view of the film. | 0 | + | 0 | 0 | + | + | 0 | + | 0 | + | + |
| thread #51 w 0.139 The escalation from Billy not trusting Elizabeth to properly whisk eggs to him creating an AI model of her and taking power of attorney was horrifying VADER (old v1): negative — VADER -0.58 siebert+opinion (LIVE): negative — P(op)=0.86 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.66 GPT-6.1 Sol: no_verdict — Describes a horrifying plot escalation without expressing approval or criticism of the film itself. Claude Opus 5.5: no_verdict — Describes a character's horrifying escalation in the plot without judging the film's quality. Claude Sonnet 5.5: no_verdict — Describes plot escalation as horrifying, but that's about the character's actions, not a judgement of the film's quality. Gemini 4 Argon: no_verdict — Describes the horrifying escalation of plot events without expressing an opinion on the quality of the film. Grok 4.7: no_verdict — Describes a horrifying plot escalation without judging the film. GPT-6 Luna (cheap): no_verdict — Describes the disturbing escalation but gives no opinion on the film itself. Claude Haiku 5.5 (cheap): no_verdict — Describes a plot escalation as horrifying but gives no judgement of the film's quality or whether they'd recommend it. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment describes a horrifying plot point without expressing a judgment or recommendation about the film itself. | − | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #52 w 0.137 Of the million things that got lost in the madness of the film, the various reptiles skittering around his Austin complex were some of my favorites VADER (old v1): negative — VADER -0.34 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=0.99 Local NLI stance (deberta-v3-large): negative — NLI p=0.65 GPT-6.1 Sol: positive — Expresses fondness for the film’s reptile details without clear criticism of the film. Claude Opus 5.5: positive — Affectionately calls the film's details, like the skittering reptiles, among their favorites; 'madness' reads as fond, not critical. Claude Sonnet 5.5: mixed — Calls the film 'madness' where things got lost (criticism), but affectionately singles out the reptiles as favorites. Gemini 4 Argon: positive — The commenter fondly highlights a favorite detail amidst the film's madness, indicating enjoyment without actual criticism. Grok 4.7: positive — Fondly recalls favorite details amid the film's chaos. GPT-6 Luna (cheap): positive — They describe the reptiles as some of their favorite parts of the film. Claude Haiku 5.5 (cheap): mixed — Calls the film 'madness' where things got lost, but singles out reptiles as favorites; sarcastic-leaning criticism with a small positive note. Gemini 3.5 Flash-Lite (cheap): positive — The commenter highlights specific elements of the film as their favorites, indicating enjoyment. | − | + | − | + | + | 0 | + | + | + | 0 | + |
| thread #53 w 0.136 owning exotic pets is a huge telltale sign if narcissism....I'll just leave this here VADER (old v1): positive — VADER 0.273 siebert+opinion (LIVE): negative — P(op)=0.43 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.48 GPT-6.1 Sol: no_verdict — Comments on exotic pet ownership and narcissism without judging the film. Claude Opus 5.5: no_verdict — Comment about a character trait or real-world observation; no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Comment makes a general remark about exotic pet owners and narcissism, with no judgement of the film itself. Gemini 4 Argon: no_verdict — The comment makes an observation about a character or plot detail without expressing an opinion on the film itself. Grok 4.7: no_verdict — Character observation about narcissism, no judgment of the film. GPT-6 Luna (cheap): no_verdict — Comments on exotic pet ownership and narcissism, without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Comment is a general remark about narcissism and exotic pets, with no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses owning exotic pets and narcissism without judging the film. | + | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #54 w 0.135 Getting Elizabeth to watch Amanda in The Dropout was great. Getting Amanda to watch Elizabeth watching Amanda in The Dropout was brilliant. Seyfried-ception VADER (old v1): positive — VADER 0.836 siebert+opinion (LIVE): positive — P(op)=0.65 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.83 GPT-6.1 Sol: no_verdict — Praises a meta viewing stunt involving The Dropout, without judging the film itself. Claude Opus 5.5: positive — Calls the film's meta casting idea 'brilliant', praising the movie's concept. Claude Sonnet 5.5: no_verdict — Comment praises casting/meta aspect of The Dropout and Seyfried, not a verdict on the film 'You Can See Everything' itself. Gemini 4 Argon: positive — The commenter praises specific moments in the film as 'great' and 'brilliant'. Grok 4.7: no_verdict — Joke about The Dropout, not a judgment of this film. GPT-6 Luna (cheap): positive — Calls the film’s meta casting conceit great and brilliant. Claude Haiku 5.5 (cheap): no_verdict — Joke about casting and Seyfried's roles in other projects; no judgement of this film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment makes a joke about the casting and meta-reference without judging the film itself. | + | + | 0 | 0 | + | 0 | + | 0 | + | 0 | 0 |
| thread #55 w 0.134 Making Amanda Seyfried pronounce Seyfried incorrectly 🤯😂 VADER (old v1): positive — VADER 0.637 siebert+opinion (LIVE): neutral — P(op)=0.21 P(pos)=0.92 Local NLI stance (deberta-v3-large): mixed — NLI p=0.40 GPT-6.1 Sol: no_verdict — Jokes about Amanda Seyfried’s pronunciation without judging the film. Claude Opus 5.5: no_verdict — Joke about name pronunciation; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — A joke about the pronunciation of an actress's name, with no judgement of the film's quality or any recommendation. Gemini 4 Argon: no_verdict — The comment makes a joke/observation about a specific line or trivia without giving an overall opinion on the film. Grok 4.7: no_verdict — Joke about name pronunciation, no film judgment. GPT-6 Luna (cheap): no_verdict — A joke about pronunciation, with no judgment of the film itself. Claude Haiku 5.5 (cheap): no_verdict — Joke about the actress's name pronunciation; no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment makes a joke about a pronunciation error without judging the film itself. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #56 w 0.132 I’ve never had such a visceral reaction to someone not getting a joke. This was unlike anything I have ever seen before and I don’t think I’ll stop chewing on it anytime soon. There is something so unsettlingly beautiful about the whole experience that is consistently… VADER (old v1): positive — VADER 0.915 siebert+opinion (LIVE): positive — P(op)=0.94 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.88 GPT-6.1 Sol: positive — Explicitly loved the film, calling it a masterpiece and praising its unique, unsettlingly beautiful viewing experience. Claude Opus 5.5: positive — Calls it unique, unsettlingly beautiful, 'fucking loved this' and a 'flat out masterpiece'. Claude Sonnet 5.5: positive — Calls it a flat-out masterpiece, says they loved it, and praises its unique, unsettlingly beautiful experience. Gemini 4 Argon: positive — The commenter explicitly states they loved the film and calls it a masterpiece. Grok 4.7: positive — Calls it a unique masterpiece they loved and keep thinking about. GPT-6 Luna (cheap): positive — They call it a masterpiece, say they loved it, and praise its unique, unsettling beauty. Claude Haiku 5.5 (cheap): positive — VADER 0.915 Gemini 3.5 Flash-Lite (cheap): positive — The commenter describes the film as a 'flat out masterpiece' and states they 'fucking loved this'. | + | + | + | + | + | + | + | + | + | + | + |
| thread #57 w 0.131 The audience literally had a stake in it. There’s a part of the movie where you’re thinking, “Wait, did I just give some money to Billy? By buying a ticket and watching?) The exact storm of events leading to that reveal makes this movie so good VADER (old v1): positive — VADER 0.625 siebert+opinion (LIVE): positive — P(op)=0.79 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.47 GPT-6.1 Sol: positive — Explicitly praises the reveal and buildup, saying they make the movie so good. Claude Opus 5.5: positive — Commenter praises the reveal's buildup and explicitly says it makes the movie so good. Claude Sonnet 5.5: positive — Commenter says the audience's involvement and the buildup to the reveal make the movie 'so good', a clear positive verdict. Gemini 4 Argon: positive — The commenter explicitly states that the sequence of events and reveal "makes this movie so good." Grok 4.7: positive — Praises the reveal and says the movie is so good. GPT-6 Luna (cheap): positive — They explicitly praise the reveal and say the events leading to it make the movie so good. Claude Haiku 5.5 (cheap): positive — Commenter praises a plot reveal and its engagement with the audience, saying it makes the movie 'so good'. Gemini 3.5 Flash-Lite (cheap): positive — The commenter explicitly states that the storm of events makes this movie so good. | + | + | + | + | + | + | + | + | + | + | + |
| thread #58 w 0.13 Definitley! I was relieved when the card came up at the end that said he changed his mind about the back end. VADER (old v1): positive — VADER 0.439 siebert+opinion (LIVE): positive — P(op)=0.63 P(pos)=0.87 Local NLI stance (deberta-v3-large): mixed — NLI p=0.33 GPT-6.1 Sol: no_verdict — Expresses relief about an ending revelation, without judging the film’s quality or recommending it. Claude Opus 5.5: no_verdict — Comments on an ending plot detail and feeling relieved; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Comment discusses a plot detail and the commenter's relief about it, without giving an overall judgement of the film's quality. Gemini 4 Argon: no_verdict — Discusses a plot point at the end of the film without expressing a verdict on the movie itself. Grok 4.7: no_verdict — Agrees about an ending plot detail; no film judgement. GPT-6 Luna (cheap): no_verdict — They express relief about a plot detail in the ending card, but don’t judge the film itself. Claude Haiku 5.5 (cheap): no_verdict — Comments on a plot detail (an ending card about a character changing his mind) without judging the film's quality or recommending it. Gemini 3.5 Flash-Lite (cheap): positive — The commenter expresses clear relief and satisfaction with the film's resolution. | + | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | + |
| thread #59 w 0.129 The "Who's On First" scene just broke my brain so much. VADER (old v1): negative — VADER -0.421 siebert+opinion (LIVE): negative — P(op)=0.99 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.55 GPT-6.1 Sol: no_verdict — Expresses bewilderment at a specific scene without clearly praising or criticizing the film. Claude Opus 5.5: no_verdict — Strong reaction to one scene, but 'broke my brain' is ambiguous and gives no clear judgement of the film. Claude Sonnet 5.5: no_verdict — Says a scene 'broke my brain' but it's ambiguous whether praise or criticism; no clear judgment of the film overall. Gemini 4 Argon: no_verdict — The comment reacts to the complexity of a specific scene without expressing whether they liked or disliked the film. Grok 4.7: no_verdict — Scene reaction only; no clear overall film judgement. GPT-6 Luna (cheap): no_verdict — They react to a scene but don't clearly judge the film or recommend it. Claude Haiku 5.5 (cheap): positive — The commenter says a scene 'just broke my brain', an enthusiastic reaction implying they were impressed by it. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses a specific scene without expressing a clear verdict on the film. | − | − | − | 0 | 0 | 0 | 0 | 0 | 0 | + | 0 |
| thread #60 w 0.128 The most genius idea to present to them on Nathan’s behalf, and that she didn’t get the premise of the joke was so so so baffling. But you definitely did get the sense in that moment that she was deeply embarrassed and didn’t want to show it. VADER (old v1): positive — VADER 0.229 siebert+opinion (LIVE): negative — P(op)=0.90 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.51 GPT-6.1 Sol: no_verdict — Discusses a joke and a character’s embarrassment without expressing an overall judgment of the film. Claude Opus 5.5: no_verdict — Discusses a character's reaction in a scene; no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Discusses a specific scene and a character's reaction; calling an idea 'genius' is about in-story content, not an overall verdict on the film. Gemini 4 Argon: positive — Praises a concept in the show/film as 'the most genius idea' while discussing a character's reaction. Grok 4.7: no_verdict — Scene and character reaction discussion, no film judgement. GPT-6 Luna (cheap): no_verdict — Discusses a character’s reaction to a joke, without judging the film itself. Claude Haiku 5.5 (cheap): no_verdict — Discusses a plot moment and a character's reaction to a joke, with no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — Discusses plot and character reactions without expressing a judgement or recommendation on the film itself. | + | − | 0 | 0 | 0 | 0 | + | 0 | 0 | 0 | 0 |
| thread #61 w 0.127 It's so crazy because I just really don't know if she truly doesn't get or if she's SO committed to the bit that she won't relent. VADER (old v1): positive — VADER 0.595 siebert+opinion (LIVE): negative — P(op)=0.86 P(pos)=0.01 Local NLI stance (deberta-v3-large): mixed — NLI p=0.45 GPT-6.1 Sol: no_verdict — Speculates about someone's understanding or commitment to a bit without judging the film. Claude Opus 5.5: no_verdict — Speculates about a character's or person's motives without judging the film's quality. Claude Sonnet 5.5: no_verdict — Speculates about a character's motives (whether she truly doesn't get it or is committed to the bit) without judging the film itself. Gemini 4 Argon: no_verdict — Discusses a character's or subject's behavior without expressing an opinion on the film's quality. Grok 4.7: no_verdict — Speculates about a character's motives, no film judgement. GPT-6 Luna (cheap): no_verdict — Speculates about a character's intentions without expressing an opinion on the film. Claude Haiku 5.5 (cheap): no_verdict — Speculates about a character's motivation or intent; no judgement of the film's quality or recommendation. Gemini 3.5 Flash-Lite (cheap): no_verdict — Discusses a character's behavior without expressing an opinion or verdict on the film itself. | + | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #62 w 0.126 The audience at my showing erupted in laughter when dude brought up Charlie Kirk to Amanda. Of course that guy loved Charlie VADER (old v1): positive — VADER 0.796 siebert+opinion (LIVE): neutral — P(op)=0.25 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.36 GPT-6.1 Sol: no_verdict — Describes audience laughter and jokes about a character’s politics without evaluating the film. Claude Opus 5.5: no_verdict — Describes audience reaction and a character detail; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Describes audience reaction and a plot moment with a joke about a character; no explicit judgement of the film itself. Gemini 4 Argon: no_verdict — Describes an audience reaction to a specific scene and character detail without judging the overall film. Grok 4.7: no_verdict — Notes audience laughter at a scene; no judgment of the film. GPT-6 Luna (cheap): no_verdict — Describes the audience’s reaction to a scene and a character’s politics, without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Describes audience laughter at a plot moment and a jab at a character's politics; no clear judgement of the film's quality. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses audience reaction and character traits without judging the film itself. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #63 w 0.125 Of course he would dm Elon. Of course he’s an Ai bro. Of course he’s definitely an emotional abuser/possible domestic one too. Just from what she said about how much he takes care of her and they would build a monument to their love. No fucking wonder she gave up her power of… VADER (old v1): positive — VADER 0.859 siebert+opinion (LIVE): negative — P(op)=0.42 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.86 GPT-6.1 Sol: no_verdict — Discusses a person's behavior and relationship dynamics without evaluating the film itself. Claude Opus 5.5: no_verdict — Comment criticizes a person/character's behavior, not the film itself; no judgement of the movie. Claude Sonnet 5.5: no_verdict — Comment criticizes a character's behavior and discusses plot details without giving any judgment of the film itself. Gemini 4 Argon: no_verdict — The comment only criticizes and discusses a character's actions and personality without judging the film itself. Grok 4.7: no_verdict — Character and plot speculation only; no judgment of the film. GPT-6 Luna (cheap): no_verdict — Criticizes and speculates about a character’s behavior, but gives no opinion on the film itself. Claude Haiku 5.5 (cheap): no_verdict — Commenter discusses a character's behavior and speculates about his abuse; no judgement of the film itself is expressed. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses the characters and plot events without giving a verdict on the film itself. | + | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #64 w 0.124 Incredible film from beginning to end. 10/10 One thing that I haven’t seen discussed is the dark turn Billy took in that 2 year gap. He was always problematic but in that gap he became totally radicalized by right wing propoganda on Twitter like half of the men in this country.… VADER (old v1): positive — VADER 0.519 siebert+opinion (LIVE): positive — P(op)=0.98 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.52 GPT-6.1 Sol: positive — Explicitly calls it an incredible film from beginning to end and rates it 10/10. Claude Opus 5.5: positive — Calls it an incredible film, 10/10; the character criticism is about Billy, not the film. Claude Sonnet 5.5: positive — Commenter calls it an incredible film from beginning to end and gives it 10/10; the rest is analysis of a character. Gemini 4 Argon: positive — The commenter explicitly rates the movie 10/10 and calls it an 'Incredible film from beginning to end.' Grok 4.7: positive — Calls the film incredible and rates it 10/10. GPT-6 Luna (cheap): positive — Calls it an incredible film from beginning to end and rates it 10/10. Claude Haiku 5.5 (cheap): positive — Explicitly calls the film 'incredible' and gives it 10/10; the rest is interpretation of the plot and characters. Gemini 3.5 Flash-Lite (cheap): positive — The commenter calls the movie an "Incredible film from beginning to end" and gives it a "10/10" rating. | + | + | + | + | + | + | + | + | + | + | + |
| thread #65 w 0.123 His face also had the alcohol bloat. VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): negative — P(op)=0.49 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.35 GPT-6.1 Sol: no_verdict — Describes someone's appearance without judging the film. Claude Opus 5.5: no_verdict — Observation about a character's appearance; no judgement of the film. Claude Sonnet 5.5: no_verdict — Observation about an actor's appearance; no judgement of the film itself. Gemini 4 Argon: no_verdict — An observation about a character or actor's appearance with no judgment on the film itself. Grok 4.7: no_verdict — Observes a character's appearance; no film judgement. GPT-6 Luna (cheap): no_verdict — A descriptive observation about someone's appearance, with no opinion on the film. Claude Haiku 5.5 (cheap): no_verdict — Describes an actor's appearance from the film with no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses a character's appearance without expressing an opinion on the film. | 0 | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #66 w 0.122 It’s hard to really wrangle my thoughts on this one. I just came out of the theater an hour ago and there is just so much to think and talk about. I think Holmes is an incredible documentary subject and the summation of my thoughts is we are lucky to have this, even though I… VADER (old v1): positive — VADER 0.998 siebert+opinion (LIVE): neutral — P(op)=0.07 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.96 GPT-6.1 Sol: positive — Calls the documentary special and says we are lucky to have it, praising its fascinating subject and unanswered questions. Claude Opus 5.5: positive — Calls Holmes an incredible subject and says 'we are lucky to have this'; the questions it leaves open are framed as fascinating. Claude Sonnet 5.5: positive — Calls the documentary subject incredible and says 'we are lucky to have this', praising its fascination despite unanswered questions. Gemini 4 Argon: positive — The commenter calls the film 'special' and states 'we are lucky to have this.' Grok 4.7: positive — Calls the film special and says we are lucky to have it. GPT-6 Luna (cheap): positive — Calls it an incredible, special documentary and says viewers are lucky to have it, despite unanswered questions. Claude Haiku 5.5 (cheap): positive — VADER 0.998 Gemini 3.5 Flash-Lite (cheap): positive — The commenter finds the documentary fascinating and states 'we are lucky to have this'. | + | 0 | 0 | + | + | + | + | + | + | + | + |
| thread #67 w 0.121 I’m not sure I believe that she didn’t understand the joke on “who’s on first.” I think she’s smart enough to know that it plays in her favor to appear as if she doesn’t understand the joke. But I think it goes back to what’s so frustrating about all this, is that there’s never… VADER (old v1): negative — VADER -0.294 siebert+opinion (LIVE): neutral — P(op)=0.20 P(pos)=0.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.67 GPT-6.1 Sol: no_verdict — Discusses the subject’s understanding and motives, without evaluating the film itself. Claude Opus 5.5: no_verdict — Discusses a character's possible deception and ambiguity; frustration is about the situation, not a judgment of the film. Claude Sonnet 5.5: no_verdict — Discusses a character's motives and ambiguity in a scene; no judgment of the film's quality or recommendation. Gemini 4 Argon: no_verdict — The comment discusses a specific scene and character motivation without offering a judgment on the film itself. Grok 4.7: no_verdict — Character and plot speculation only; no judgment of the film. GPT-6 Luna (cheap): negative — They describe the film’s unresolved ambiguity as frustrating, signaling dissatisfaction with how it handles the subject. Claude Haiku 5.5 (cheap): no_verdict — VADER -0.294 Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses character motivations and plot ambiguities without expressing a judgment on the film. | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | − | 0 | 0 |
| thread #68 w 0.12 Such a phenomenal movie. My favorite part I don’t see a lot of people discussing is Billy watching biology videos on YouTube at 3x speed. Him having apoptosis (a concept I’ve taught 8th graders) explained to him is hysterical. VADER (old v1): positive — VADER 0.44 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.77 GPT-6.1 Sol: positive — Calls the movie phenomenal and praises the biology video scene as hysterical. Claude Opus 5.5: positive — Calls it a phenomenal movie and enthusiastically praises a favorite humorous scene. Claude Sonnet 5.5: positive — Calls it a 'phenomenal movie' and enthusiastically praises a favorite scene as hysterical. Gemini 4 Argon: positive — The commenter explicitly calls it a "phenomenal movie" and praises a humorous scene. Grok 4.7: positive — Calls the movie phenomenal and praises a favorite scene. GPT-6 Luna (cheap): positive — Calls the movie phenomenal and praises a specific scene as hysterical. Claude Haiku 5.5 (cheap): positive — Calls the movie 'phenomenal' and praises a favorite scene with enthusiasm. Gemini 3.5 Flash-Lite (cheap): positive — The commenter calls the movie 'such a phenomenal movie' and shares a favorite scene. | + | + | 0 | + | + | + | + | + | + | + | + |
| thread #69 w 0.12 The moment when you realize “Wait, did I just give money to Billy, by getting a ticket to this movie? Am I getting duped?” VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): negative — P(op)=0.46 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.70 GPT-6.1 Sol: no_verdict — Jokes about ticket money benefiting Billy without expressing whether the film was good or bad. Claude Opus 5.5: no_verdict — A joke about the film's meta premise and Billy; it gives no judgement of the film's quality. Claude Sonnet 5.5: negative — Sarcastic remark implying the viewer feels duped into paying for the movie, a negative judgment of the film/its marketing. Gemini 4 Argon: no_verdict — The comment makes a joke about a character or subject in the film without expressing an opinion on the movie. Grok 4.7: negative — Feels possibly duped after paying for a ticket. GPT-6 Luna (cheap): negative — They regret paying for a ticket and feel the movie may have duped them. Claude Haiku 5.5 (cheap): negative — Commenter jokes about feeling duped by buying a ticket, implying the film was a disappointment or waste of money. Gemini 3.5 Flash-Lite (cheap): negative — The commenter feels cheated and regrets paying for the movie, indicating a negative verdict. | 0 | − | − | 0 | 0 | − | 0 | − | − | − | − |
| thread #70 w 0.119 I think my theater breathed a collective sigh of relief at the end lol VADER (old v1): positive — VADER 0.718 siebert+opinion (LIVE): positive — P(op)=0.74 P(pos)=1.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.43 GPT-6.1 Sol: no_verdict — Describes the audience’s relief at the ending without clearly evaluating the film. Claude Opus 5.5: no_verdict — Describes the audience's relief at the ending; it's unclear whether that reflects tension released or boredom, so no clear judgement. Claude Sonnet 5.5: no_verdict — Describes the audience's reaction at the end (tension relief) without stating whether they liked the film or would recommend it. Gemini 4 Argon: no_verdict — Describes an ambiguous audience reaction at the ending without stating whether the film was good or bad. Grok 4.7: no_verdict — Notes audience relief at the ending, no clear like or dislike. GPT-6 Luna (cheap): no_verdict — A collective sigh of relief describes the audience’s reaction, but doesn’t clearly judge the film. Claude Haiku 5.5 (cheap): no_verdict — Ambiguous joke about the ending's effect on the audience; no clear judgement of the film's quality or recommendation. Gemini 3.5 Flash-Lite (cheap): positive — Audience relief suggests a tense, impactful viewing experience typical of thrillers or horror films. | + | + | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | + |
| thread #71 w 0.118 Amanda was the perfect stand in for the audience VADER (old v1): positive — VADER 0.572 siebert+opinion (LIVE): neutral — P(op)=0.38 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.47 GPT-6.1 Sol: positive — Praises Amanda as a perfectly effective audience surrogate. Claude Opus 5.5: positive — Praises the character's effectiveness as an audience surrogate, a favorable comment on the film's craft. Claude Sonnet 5.5: positive — Praises the character Amanda as a perfect audience surrogate, implying approval of the film's writing or craft, though only mildly stated. Gemini 4 Argon: positive — Praising a character as a 'perfect stand in for the audience' compliments the film's writing and execution. Grok 4.7: no_verdict — Character observation only; no judgment of the film. GPT-6 Luna (cheap): positive — Calling Amanda the perfect audience stand-in is praise for the film's character and perspective. Claude Haiku 5.5 (cheap): no_verdict — Describes the character's role as audience stand-in without any judgement of whether the film is good or recommended. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses a character's role without expressing a judgment or recommendation for the film. | + | 0 | 0 | + | + | + | + | 0 | + | 0 | 0 |
| thread #72 w 0.117 Amanda was great but she was also just another problem to this whole thing She got mad at Billy for portraying Liz and then the next minute defended herself portraying Liz because it was "all a character" She was no hero. Sure, she says Billy sucks which is 100% true, but… VADER (old v1): negative — VADER -0.933 siebert+opinion (LIVE): negative — P(op)=0.94 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.92 GPT-6.1 Sol: no_verdict — Discusses Amanda’s hypocrisy and responsibility within the story, without judging the film itself. Claude Opus 5.5: no_verdict — Analyzes Amanda's character hypocrisy; 'great' refers to character role, no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Discusses characters' behavior and morality (Amanda, Billy); 'Amanda was great' is a passing remark, with no clear verdict on the film itself. Gemini 4 Argon: no_verdict — The comment only discusses and critiques the actions of the people/characters involved, offering no opinion on the film itself. Grok 4.7: no_verdict — Criticizes a character's hypocrisy, not the film itself. GPT-6 Luna (cheap): no_verdict — The comment critiques Amanda’s character and actions, without judging the film itself. Claude Haiku 5.5 (cheap): no_verdict — Critiques a character's behavior and real-life conduct, not the film's quality or whether they'd recommend it. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses character actions and morality without giving a verdict on the film itself. | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #73 w 0.116 I do think it’s very different for a known actor to portray someone in a show than it is to tweet from the real life persons account as if you are them. However, something I found really interesting is Amanda’s argument that it was not her fault that people didn’t know it was a… VADER (old v1): positive — VADER 0.375 siebert+opinion (LIVE): neutral — P(op)=0.19 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.66 GPT-6.1 Sol: no_verdict — Discusses Amanda’s ethical arguments and involvement without judging the film’s quality or recommending it. Claude Opus 5.5: no_verdict — Discusses Amanda's argument and a comparison to Zuckerberg; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Discusses themes and the subject's arguments, with curiosity about her view, but gives no judgement of the film's quality or a recommendation. Gemini 4 Argon: no_verdict — Discusses the themes and subjects of the film without expressing a judgment on the movie's quality. Grok 4.7: no_verdict — Discusses character ethics and curiosity, no film judgement. GPT-6 Luna (cheap): no_verdict — Discusses Amanda’s argument and curiosity about her views, but gives no verdict on the film. Claude Haiku 5.5 (cheap): no_verdict — Discusses the film's themes and a character's argument, plus curiosity about the director's view, without judging whether the film is good or recommended. Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter discusses themes, comparisons, and character motivations without expressing a judgment or verdict on the film itself. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #74 w 0.115 My immediate takeaway is that this movie exists because of the "Drink my grandsons pee" guy. VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): negative — P(op)=0.98 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.93 GPT-6.1 Sol: no_verdict — The comment jokes about the movie’s inspiration without expressing whether the commenter liked or disliked it. Claude Opus 5.5: no_verdict — Joke about the movie's origin or inspiration; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Comment jokes about the film's origin/inspiration without stating whether they liked or disliked the film itself. Gemini 4 Argon: no_verdict — The comment makes a joke/observation about the movie's inspiration without expressing an opinion on its quality. Grok 4.7: no_verdict — Joke about the film's origin, no judgment of quality. GPT-6 Luna (cheap): no_verdict — A joking observation about why the movie exists, without judging whether it is good or worth watching. Claude Haiku 5.5 (cheap): no_verdict — Joking about the film's origin via a viral meme; no judgement of the film's quality or whether to watch it. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment makes a joke and a reference without expressing a verdict on the film. | 0 | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #75 w 0.115 Even in my critically acclaimed documentary biopics I can’t escape Charlie Kirk VADER (old v1): positive — VADER 0.178 siebert+opinion (LIVE): negative — P(op)=0.76 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.36 GPT-6.1 Sol: no_verdict — Jokes about Charlie Kirk appearing in documentary biopics without expressing a clear opinion of the film. Claude Opus 5.5: no_verdict — Joke about Charlie Kirk appearing; no judgement of the film's quality itself. Claude Sonnet 5.5: no_verdict — A joke about Charlie Kirk appearing everywhere; offers no judgement of the film's quality or a recommendation. Gemini 4 Argon: no_verdict — The comment makes a joke about a figure's appearance in the film without expressing an opinion on the movie itself. Grok 4.7: no_verdict — Joke about Charlie Kirk, no judgement of the film. GPT-6 Luna (cheap): no_verdict — A joke about Charlie Kirk appearing in documentaries, with no clear judgment of this film. Claude Haiku 5.5 (cheap): no_verdict — Sarcastic joke about Charlie Kirk appearing in documentaries; no judgement of the film itself is expressed. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment is a joke/reference unrelated to judging the film itself. | + | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #76 w 0.114 Why did you take over the eggs? VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): neutral — P(op)=0.07 P(pos)=0.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.47 GPT-6.1 Sol: no_verdict — The comment asks a question without expressing any judgment of the film. Claude Opus 5.5: no_verdict — A question about a plot detail, with no judgement of the film. Claude Sonnet 5.5: no_verdict — A question, likely about a plot detail or quote, with no judgement of the film. Gemini 4 Argon: no_verdict — The comment is just a quote or plot question without expressing any opinion on the film. Grok 4.7: no_verdict — Plot question or quote with no judgment of the film. GPT-6 Luna (cheap): no_verdict — A question about the eggs, with no opinion about the film. Claude Haiku 5.5 (cheap): no_verdict — A plot-related question about a scene; no judgement of the film expressed. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment is just a plot quote or question without any judgment of the film. | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #77 w 0.113 he clocked billy immediately. it's funny billy and elizabeth wanted nathan because they viewed this as a human interest story (aka they wanted to get people to emotionally connect to elizabeth and thought nathan would pull that out) but they underestimated how laser (no pun… VADER (old v1): positive — VADER 0.747 siebert+opinion (LIVE): neutral — P(op)=0.04 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.84 GPT-6.1 Sol: no_verdict — Discusses characters’ motivations and social dynamics without evaluating the film. Claude Opus 5.5: no_verdict — Character and plot analysis only; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Plot and character analysis about Nathan, Billy and Elizabeth's dynamics, with no judgement of the film's quality or recommendation. Gemini 4 Argon: no_verdict — Discusses character dynamics and plot details without expressing an opinion on the quality of the film itself. Grok 4.7: no_verdict — Plot analysis of character motives; no judgment of the film. GPT-6 Luna (cheap): no_verdict — Discusses character motivations and social dynamics without judging the film itself. Claude Haiku 5.5 (cheap): no_verdict — Analyzes plot and character motivations in the story without judging whether the film is good or worth recommending. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses plot and character dynamics without expressing any judgment or recommendation of the film. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #78 w 0.113 This is one of Nathan’s superpowers. He does not feel awkwardness. It bounces right off him. He’ll ask the questions that social politeness usually censors. He’ll stay on the point, maintain the eye contact, let the silence hold as long as you want it to. VADER (old v1): negative — VADER -0.098 siebert+opinion (LIVE): neutral — P(op)=0.01 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.70 GPT-6.1 Sol: no_verdict — Describes Nathan’s personality and conversational behavior without judging the film. Claude Opus 5.5: no_verdict — Describes Nathan's personality traits without evaluating the film itself. Claude Sonnet 5.5: no_verdict — Describes a person's (Nathan's) character traits without giving any judgment of the film itself. Gemini 4 Argon: no_verdict — The comment describes the subject's interview style and personality without offering a verdict on the film itself. Grok 4.7: no_verdict — Character description only; no judgment of the film. GPT-6 Luna (cheap): no_verdict — Describes Nathan’s social behavior without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Describes a character's traits (Nathan's lack of awkwardness) without giving any judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses a character's traits without expressing any opinion or judgment about the film itself. | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #79 w 0.112 Pathetic from Fielder. They do like a 5 minute scene where they discuss Who's on First and despite Elizabeth saying I don't know multiple times, he never explains to her that he's actually on third base You'd never guess he went to one of Canada's top business schools based on… VADER (old v1): negative — VADER -0.103 siebert+opinion (LIVE): negative — P(op)=1.00 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.98 GPT-6.1 Sol: no_verdict — Jokes about Fielder failing to explain a Who’s on First reference, without genuinely evaluating the film. Claude Opus 5.5: no_verdict — Joking riff on the Who's on First scene and Fielder's character; no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Criticizes a character's behavior within a scene, joking about the character; no clear judgment of the film itself. Gemini 4 Argon: no_verdict — The comment is a joke referencing the Abbott and Costello routine and Nathan Fielder's catchphrase, without judging the film. Grok 4.7: negative — Calls the scene pathetic and criticizes its execution. GPT-6 Luna (cheap): negative — Calls the scene pathetic and criticizes Fielder’s handling of it. Claude Haiku 5.5 (cheap): negative — VADER -0.103 Gemini 3.5 Flash-Lite (cheap): negative — Criticizes the scene and calls the creator's work pathetic. | − | − | − | 0 | 0 | 0 | 0 | − | − | − | − |
| thread #80 w 0.111 Well, pretty sure I love Amanda Seyfried as much as I hate Billy Evans now. VADER (old v1): positive — VADER 0.796 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.82 GPT-6.1 Sol: no_verdict — Expresses affection for Amanda Seyfried and hatred of Billy Evans, without judging the film itself. Claude Opus 5.5: positive — Loves Seyfried's performance and strongly hates the villain, implying both were effective; no criticism of the film. Claude Sonnet 5.5: no_verdict — Expresses love for the actress and hatred of a character, but gives no judgement of the film itself. Gemini 4 Argon: positive — Expresses love for the lead actress and strong emotional investment against a character/subject. Grok 4.7: no_verdict — Actor/character feelings only; no judgment of the film. GPT-6 Luna (cheap): no_verdict — They praise an actor and dislike Billy Evans, but do not judge the film itself. Claude Haiku 5.5 (cheap): no_verdict — Comment expresses feelings about actress and a character, not a verdict on the film itself. Gemini 3.5 Flash-Lite (cheap): positive — The commenter states they love the actress's performance, indicating a positive view of the film. | + | + | 0 | 0 | + | 0 | + | 0 | 0 | 0 | + |
| thread #81 w 0.11 “He’s a fucking pussy” VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): negative — P(op)=0.98 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.70 GPT-6.1 Sol: no_verdict — A quoted insult about a character gives no judgment of the film itself. Claude Opus 5.5: no_verdict — Quote or insult about a character; no judgement of the film itself. Claude Sonnet 5.5: no_verdict — Insulting a character (likely a quote) without any judgment of the film itself. Gemini 4 Argon: no_verdict — The comment is just a quote without any opinion on the film itself. Grok 4.7: no_verdict — Insult or quote about a character, no judgement of the film. GPT-6 Luna (cheap): no_verdict — A quoted insult about a character gives no judgment of the film. Claude Haiku 5.5 (cheap): no_verdict — Insult directed at a character, with no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment is just a movie quote without any verdict on the film itself. | 0 | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #82 w 0.11 Really appreciated Amanda saying out loud all the things I was thinking. VADER (old v1): positive — VADER 0.556 siebert+opinion (LIVE): neutral — P(op)=0.31 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.29 GPT-6.1 Sol: no_verdict — Appreciates Amanda voicing their thoughts, but gives no clear judgment of the film itself. Claude Opus 5.5: positive — Appreciates a character voicing their thoughts, implying approval of the film's writing, though the praise is mild. Claude Sonnet 5.5: positive — Commenter 'really appreciated' a character voicing their own thoughts, a favorable reaction to the film's writing, though fairly mild. Gemini 4 Argon: no_verdict — Expresses appreciation for a character's dialogue without giving an overall verdict on the film itself. Grok 4.7: positive — Appreciates a character voicing their own thoughts, a clear positive reaction. GPT-6 Luna (cheap): positive — They express appreciation for Amanda’s dialogue, indicating they enjoyed this aspect of the film. Claude Haiku 5.5 (cheap): positive — Commenter says they appreciated a character voicing their thoughts, an approving remark about the film's content. Gemini 3.5 Flash-Lite (cheap): positive — Appreciating the character's thoughts indicates agreement and approval of the film. | + | 0 | 0 | 0 | + | + | 0 | + | + | + | + |
| thread #83 w 0.109 That second half was quite literally Obsession in real life. During that Jail conversation, If Nathan had told me that Billy had used a one-wish willow to make Elizabeth Holmes love him more than anything, I would’ve actually believed him. VADER (old v1): positive — VADER 0.366 siebert+opinion (LIVE): positive — P(op)=0.86 P(pos)=0.98 Local NLI stance (deberta-v3-large): positive — NLI p=0.37 GPT-6.1 Sol: no_verdict — Compares events to Obsession and discusses characters without expressing approval or criticism of the film. Claude Opus 5.5: no_verdict — Compares second half to the film Obsession and jokes about the jail conversation; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Compares the second half to the film 'Obsession' with a joke-like hypothetical; no clear judgment of the film's quality. Gemini 4 Argon: no_verdict — The comment makes a joke comparing the plot and characters to another work without judging the film's quality. Grok 4.7: no_verdict — Plot joke about the jail scene; no film judgement. GPT-6 Luna (cheap): no_verdict — The comment jokes about the plot resembling real-life obsession but gives no clear opinion of the film. Claude Haiku 5.5 (cheap): no_verdict — Comments on plot and a comparison to a real-life story, with no judgement of whether the film is good or worth watching. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses plot events and makes a joke comparison without judging the film itself. | + | + | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #84 w 0.108 I still don’t know if she’s the victim or the victimizer, but her taste in men is atrocious VADER (old v1): negative — VADER -0.351 siebert+opinion (LIVE): negative — P(op)=0.82 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.39 GPT-6.1 Sol: no_verdict — Discusses the character’s role and taste in men without judging the film. Claude Opus 5.5: no_verdict — Comments on character ambiguity and a joke about her taste in men; no judgement of the film. Claude Sonnet 5.5: no_verdict — Comment discusses character ambiguity and joke about her taste in men; no judgment of the film's quality or recommendation. Gemini 4 Argon: no_verdict — The comment discusses character traits and plot ambiguity without expressing an opinion on the film's quality. Grok 4.7: no_verdict — Character ambiguity and taste remark; no film judgement. GPT-6 Luna (cheap): no_verdict — Comments on the character's role and romantic choices, not whether the commenter liked the film. Claude Haiku 5.5 (cheap): no_verdict — Comments on characters' ambiguity and a character's dating choices, with no judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses character morality and plot ambiguity without expressing an opinion on the film itself. | − | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #85 w 0.108 This was masterfully shot. I was in awe the entire time. I’d say a documentary has no business looking this nice, but everyone’s faces, their expressions, eyes…it’s all so important. VADER (old v1): positive — VADER 0.554 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.44 GPT-6.1 Sol: positive — Calls the film masterfully shot and expresses sustained awe, praising its visuals and their importance. Claude Opus 5.5: positive — Praises masterful cinematography and says they were in awe the entire time. Claude Sonnet 5.5: positive — Commenter praises the film's cinematography as masterful and says they were in awe throughout, a clear positive verdict. Gemini 4 Argon: positive — The commenter strongly praises the cinematography, stating it was masterfully shot and they were in awe the entire time. Grok 4.7: positive — Praises masterful cinematography and being in awe throughout. GPT-6 Luna (cheap): positive — They express sustained awe and praise the film’s cinematography and attention to faces. Claude Haiku 5.5 (cheap): positive — Praises the film's masterful shooting and says they were in awe the entire time. Gemini 3.5 Flash-Lite (cheap): positive — Praises the film as masterfully shot and states they were in awe the entire time. | + | + | 0 | + | + | + | + | + | + | + | + |
| thread #86 w 0.107 No idea if she's eligible but give Amanda Seyfried the best supporting actress Oscar VADER (old v1): positive — VADER 0.876 siebert+opinion (LIVE): positive — P(op)=0.87 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.57 GPT-6.1 Sol: positive — Calls Amanda Seyfried's performance Oscar-worthy, expressing strong praise for the film's acting. Claude Opus 5.5: positive — Calls for an Oscar for Seyfried's performance, implying strong praise for the film's acting. Claude Sonnet 5.5: positive — Calls for Amanda Seyfried to win best supporting actress, implying strong praise for her performance and the film. Gemini 4 Argon: positive — High praise for an actor's performance in the film, suggesting they deserve an Oscar. Grok 4.7: positive — Oscar praise for Seyfried's performance endorses the film. GPT-6 Luna (cheap): positive — Urges an Oscar for Amanda Seyfried, expressing strong praise for her performance and the film. Claude Haiku 5.5 (cheap): positive — Commenter urges an Oscar for Amanda Seyfried's performance, implying strong approval of her work in the film. Gemini 3.5 Flash-Lite (cheap): positive — Praising Amanda Seyfried's performance to Oscar-worthy levels implies a strong recommendation of the film. | + | + | 0 | + | + | + | + | + | + | + | + |
| thread #87 w 0.107 It was very odd to me that Liz called Billy at home from prison while Amanda was over. She 100% knew and knew it would fuck with her head. Amanda was physically unsettled VADER (old v1): negative — VADER -0.893 siebert+opinion (LIVE): neutral — P(op)=0.39 P(pos)=0.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.43 GPT-6.1 Sol: no_verdict — Discusses Liz’s intentions and Amanda’s reaction without judging the film’s quality. Claude Opus 5.5: no_verdict — Discusses a character's manipulative plot action; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — The comment analyzes a plot point and character motivations without expressing any judgment about the film's quality or recommending it. Gemini 4 Argon: no_verdict — The comment only discusses character actions and plot events without evaluating the quality of the film. Grok 4.7: no_verdict — Plot observation only; no judgement of the film. GPT-6 Luna (cheap): no_verdict — Describes a character interaction and its effect on Amanda, without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Plot discussion about a scene and characters' motivations, with no judgement of whether the film is good or worth watching. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment only discusses plot details and character actions without expressing a verdict on the film. | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #88 w 0.106 Holmes even said like "I don't know if I knew you'd be there" and Amanda calls it out as really weird. You're definitely right there VADER (old v1): positive — VADER 0.696 siebert+opinion (LIVE): neutral — P(op)=0.34 P(pos)=0.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.57 GPT-6.1 Sol: no_verdict — Discusses character dialogue and agrees with another interpretation without evaluating the film. Claude Opus 5.5: no_verdict — Discusses a plot detail and agrees with another commenter; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Discusses a plot detail and agrees with another commenter's observation; no judgment of the film's quality or recommendation. Gemini 4 Argon: no_verdict — The comment merely discusses plot details and character dialogue without expressing an opinion on the film's quality. Grok 4.7: no_verdict — Quotes dialogue and agrees a line is weird; no film judgement. GPT-6 Luna (cheap): no_verdict — They discuss a character interaction and agree with an interpretation, without judging the film. Claude Haiku 5.5 (cheap): no_verdict — Discusses a plot moment and dialogue exchange with a reply, with no judgement of the film's quality or recommendation. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses plot details and character dialogue without expressing an opinion on the film. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #89 w 0.105 The key scene to me was when Nathan mentioned bringing up the David Blaine magic trick to Liz in their first meeting. The ice pick thru the hand. In my opinion, that trick very closes resembles what is going on here. Her and Billy are both playing specific characters and trying… VADER (old v1): positive — VADER 0.957 siebert+opinion (LIVE): positive — P(op)=0.87 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.56 GPT-6.1 Sol: positive — Explicitly calls it an amazing film and rates it 10/10. Claude Opus 5.5: positive — Detailed thematic analysis ending with explicit praise: 'Amazing film 10/10'. Claude Sonnet 5.5: positive — Commenter analyzes the film's themes approvingly and explicitly concludes 'Amazing film 10/10'. Gemini 4 Argon: positive — The commenter analyzes the themes and explicitly concludes with 'Amazing film 10/10'. Grok 4.7: positive — Detailed analysis ends with clear praise: amazing film, 10/10. GPT-6 Luna (cheap): positive — The commenter explicitly calls it an amazing film and rates it 10/10. Claude Haiku 5.5 (cheap): positive — VADER 0.957 Gemini 3.5 Flash-Lite (cheap): positive — The commenter analyzes the plot and concludes with 'Amazing film 10/10'. | + | + | + | + | + | + | + | + | + | + | + |
| thread #90 w 0.105 I’m still in complete disbelief at what I just watched. I don’t think anyone will ever know how Elizabeth Holmes’ mind works but it’s fascinating seeing how she fell into the exact same dynamic she had with Sunny. God that guy is the worst and I’m so thankful he didn’t get a… VADER (old v1): positive — VADER 0.98 siebert+opinion (LIVE): negative — P(op)=0.99 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.63 GPT-6.1 Sol: positive — Calls the film fascinating, praises Nathan Fielder’s creativity, and says the experience will stay with them. Claude Opus 5.5: positive — Calls the creator one of the greatest creative minds and says the film will stick with them; hating Sunny isn't criticism. Claude Sonnet 5.5: positive — Commenter calls Nathan Fielder one of the greatest creative minds, says the film will stick with them; disliking Sunny is about a person, not the film. Gemini 4 Argon: positive — The commenter finds the film fascinating, praises the creator's genius, and says it will stick with them. Grok 4.7: positive — Finds it fascinating and lasting; praises the creator despite hating a character. GPT-6 Luna (cheap): positive — They call it fascinating, praise Nathan Fielder, and say it will stick with them for a long time. Claude Haiku 5.5 (cheap): positive — Speaker is in awe, calls it fascinating and one of the greatest, says it will stick with them; clear praise. Gemini 3.5 Flash-Lite (cheap): positive — The commenter calls Nathan Fielder one of the greatest creative minds and says the film will stick with them. | + | − | 0 | + | + | + | + | + | + | + | + |
| thread #91 w 0.104 Sunny’s nickname at Theranos was The Enforcer and it seems Elizabeth has found herself a new enforcer in Billy to do her dirty work. Mostly what I took from this is that Elizabeth seems to be a genuinely deeply codependent person who uses this codependency like a tool to keep… VADER (old v1): positive — VADER 0.681 siebert+opinion (LIVE): neutral — P(op)=0.07 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.42 GPT-6.1 Sol: no_verdict — Analyzes Elizabeth’s behavior and relationships without evaluating the documentary’s quality or recommending it. Claude Opus 5.5: no_verdict — Analyzes Elizabeth's behavior and relationships; no judgement on the documentary's quality. Claude Sonnet 5.5: no_verdict — Discusses Elizabeth Holmes's behavior and takeaways from the documentary, but offers no judgment of the film's quality or a recommendation. Gemini 4 Argon: no_verdict — The commenter analyzes the documentary's subject and their tactics without expressing any opinion on the film's quality. Grok 4.7: no_verdict — Analyzes the subject’s tactics, not whether the film works. GPT-6 Luna (cheap): no_verdict — Discusses Elizabeth Holmes and the documentary’s framing, but gives no opinion on the film itself. Claude Haiku 5.5 (cheap): no_verdict — Analyzes the documentary's subjects and plot dynamics (Theranos, Elizabeth, Billy) without stating whether they liked or recommend the film. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses the documentary's subject and plot details without evaluating the film itself. | + | 0 | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #92 w 0.104 One of my favorite moments early in the film was when Fielder was pressing Billy about Liz and his family’s wealth. Then Billy abruptly leaves saying “I’m late for my cousin’s thing…” and Nathan says “Oh I didn’t know that”. VADER (old v1): positive — VADER 0.735 siebert+opinion (LIVE): positive — P(op)=0.90 P(pos)=0.99 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.73 GPT-6.1 Sol: positive — Calling the scene one of their favorite moments expresses appreciation for the film. Claude Opus 5.5: positive — Calls a scene one of their favorite moments, implying enjoyment of the film. Claude Sonnet 5.5: no_verdict — Names a favorite moment but only describes a scene; no overall verdict on the film's quality or recommendation. Gemini 4 Argon: no_verdict — The commenter recalls a favorite moment and quote without giving an overall evaluation of the film. Grok 4.7: positive — Calls an early scene one of their favorite moments. GPT-6 Luna (cheap): positive — Calling the moment one of their favorites expresses clear praise for the film. Claude Haiku 5.5 (cheap): positive — Commenter calls one early scene 'one of my favorite moments,' expressing clear appreciation of the film. Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter discusses a specific favorite scene and quote without giving an overall judgment of the film. | + | + | 0 | + | + | 0 | 0 | + | + | + | 0 |
| thread #93 w 0.103 the reveal where his father just bluntly drops what happened with her took all the air out of the room. and then once billy tries his hardest to take control of the narrative, it gets so much stark from there. it’s only until he defends fucking charlie kirk that my audience… VADER (old v1): neutral — VADER 0.0 siebert+opinion (LIVE): negative — P(op)=0.98 P(pos)=0.01 Local NLI stance (deberta-v3-large): negative — NLI p=0.75 GPT-6.1 Sol: positive — Praises the film’s powerful, stark emotional impact, concluding with an admiring “what a movie.” Claude Opus 5.5: positive — Describes intense, gripping scenes approvingly and ends with enthusiastic 'what a movie bro.' Claude Sonnet 5.5: positive — Praises the film's impactful reveal and escalating bleakness, ending with the enthusiastic "what a movie bro"; the dark content is treated as a strength. Gemini 4 Argon: positive — Praises the film's impactful, stark emotional beats and concludes with enthusiastic approval ('what a movie bro'). Grok 4.7: positive — Praises the film's impactful reveal and stark tone; ends with 'what a movie'. GPT-6 Luna (cheap): positive — They praise the film’s tense emotional impact and appreciate the moment that brought the audience levity. Claude Haiku 5.5 (cheap): positive — Praises the reveal and escalating tension, ends with 'what a movie bro', an enthusiastic endorsement despite the sarcastic tone about the character. Gemini 3.5 Flash-Lite (cheap): positive — The commenter expresses high praise for the film's emotional impact and concludes with "what a movie bro". | 0 | − | − | + | + | + | + | + | + | + | + |
| thread #94 w 0.103 That score was haunting and very fitting. The subject matter already had me on the edge of my seat, but the music really felt perfect for an October movie. Also, it was very interesting to see get even better in her reenactment. Having never seen the Hulu show, the clips made… VADER (old v1): positive — VADER 0.96 siebert+opinion (LIVE): neutral — P(op)=0.31 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.56 GPT-6.1 Sol: positive — Praises the haunting, fitting score, engaging subject matter, and improved reenactment performance. Claude Opus 5.5: positive — Praises the haunting score, edge-of-seat tension, and the improving reenactment performance; no criticism of the film. Claude Sonnet 5.5: positive — Praises the haunting, fitting score, says the subject kept them on the edge of their seat, and finds the reenactment interesting and improving. Gemini 4 Argon: positive — Praises the haunting score, engaging subject matter, and the actor's performance in the reenactment. Grok 4.7: positive — Praises the fitting score and engaging subject matter; notes strong performance. GPT-6 Luna (cheap): positive — They praise the score, music, and reenactment, with only a mild observation about the show’s clips. Claude Haiku 5.5 (cheap): positive — Praises the score as haunting and perfect, the subject matter gripping, and the lead's performance as improving; no criticism of the film. Gemini 3.5 Flash-Lite (cheap): positive — Praises the score and considers the subject matter gripping, indicating overall approval of the film. | + | 0 | + | + | + | + | + | + | + | + | + |
| thread #95 w 0.102 the fact his only comment to nathan in the prison scene was "doesn't she look good?" AND his comments about her being "healthy" and having abs now like.........she's literally probably starving and has prison abs ok. VADER (old v1): negative — VADER -0.845 siebert+opinion (LIVE): neutral — P(op)=0.37 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.36 GPT-6.1 Sol: no_verdict — Discusses a character’s insensitive remarks in a prison scene without judging the film itself. Claude Opus 5.5: no_verdict — Comments on a character's creepy remarks in a scene; no judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Commenter discusses a character's creepy dialogue and plot details; expresses annoyance at the character, not a judgment of the film itself. Gemini 4 Argon: no_verdict — The comment reacts to a character's dialogue and behavior in a scene without judging the film itself. Grok 4.7: no_verdict — Critiques a character's dialogue in a scene, not the film. GPT-6 Luna (cheap): no_verdict — Criticizes a character’s comments about her appearance, but gives no judgment of the film itself. Claude Haiku 5.5 (cheap): no_verdict — Discusses a character's dialogue and plot details like a scene analysis, with no stated judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses plot details and character behavior without expressing a verdict on the film. | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| thread #96 w 0.102 When Amanda Seyfried joined the film and they returned to the San Diego house, i was so excited that this was really becoming The Rehearsal The Movie. VADER (old v1): positive — VADER 0.479 siebert+opinion (LIVE): neutral — P(op)=0.29 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.74 GPT-6.1 Sol: positive — Expresses excitement about the film’s direction, enthusiastically comparing it to The Rehearsal. Claude Opus 5.5: positive — Expresses excitement and enjoyment at the film's turn into 'The Rehearsal The Movie.' Claude Sonnet 5.5: no_verdict — Expresses excitement about casting and setting during production, but gives no actual verdict on the finished film itself. Gemini 4 Argon: positive — The commenter expresses excitement about the direction the film takes. Grok 4.7: positive — Excited the film became what they hoped for. GPT-6 Luna (cheap): positive — They explicitly say they were excited about the film’s direction and concept. Claude Haiku 5.5 (cheap): no_verdict — Comment compares the film to a TV show and notes casting excitement; no clear judgement of the film's quality. Gemini 3.5 Flash-Lite (cheap): no_verdict — Comment discusses plot excitement and compares it to another show without judging the film. | + | 0 | 0 | + | + | 0 | + | + | + | 0 | 0 |
| thread #98 w 0.101 Copying a response I put in another thread, but my initial reaction is that I genuinely think that this movie will go down in history, potentially as a turning point for what we expect out of documentaries. It completely turns the genre on its head and makes the audience stare… VADER (old v1): positive — VADER 0.937 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=1.00 Local NLI stance (deberta-v3-large): positive — NLI p=0.80 GPT-6.1 Sol: positive — Calls it a masterpiece and a potentially historic turning point for documentaries, praising its innovative approach. Claude Opus 5.5: positive — Calls it a masterpiece that will go down in history and turns the documentary genre on its head. Claude Sonnet 5.5: positive — Calls the film a masterpiece that will go down in history and a turning point for documentaries, with strong enthusiastic praise. Gemini 4 Argon: positive — The commenter explicitly calls the film a masterpiece and believes it will go down in history. Grok 4.7: positive — Calls the film a historical masterpiece that reinvents documentaries. GPT-6 Luna (cheap): positive — Calls it a masterpiece and a potential turning point for documentaries, praising its originality and impact. Claude Haiku 5.5 (cheap): positive — VADER 0.937 Gemini 3.5 Flash-Lite (cheap): positive — The commenter calls the movie a masterpiece and predicts it will go down in history. | + | + | + | + | + | + | + | + | + | + | + |
| thread #99 w 0.1 Amanda's pony is such a scene stealer VADER (old v1): negative — VADER -0.402 siebert+opinion (LIVE): positive — P(op)=0.99 P(pos)=1.00 Local NLI stance (deberta-v3-large): no_verdict — NLI p=0.55 GPT-6.1 Sol: positive — Calling Amanda's pony a scene stealer praises an entertaining element of the film. Claude Opus 5.5: positive — Calling the pony a 'scene stealer' praises an element of the film, suggesting enjoyment. Claude Sonnet 5.5: positive — Calls the pony a 'scene stealer', which is light praise for an element of the film, suggesting enjoyment, though not an explicit overall verdict. Gemini 4 Argon: no_verdict — The comment makes a lighthearted observation about a pony without expressing an overall opinion on the film itself. Grok 4.7: positive — Compliments the pony as a memorable scene stealer. GPT-6 Luna (cheap): positive — Calling Amanda’s pony a scene stealer expresses praise for an entertaining part of the film. Claude Haiku 5.5 (cheap): no_verdict — Comment praises a character's presence but gives no overall judgement of the film itself. Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment only mentions a character/prop without judging the film itself. | − | + | 0 | + | + | + | 0 | + | + | 0 | 0 |
| other-sub #0 w 0.5 At any point Nathan spoke to Elizabeth Holmes, she was speaking to a Judge of public opinion. Nathan mentions "I do not have a background on the science of the matter" yet he does not follow up with expert opinion, he does not examine the evidence or the claims present, nor… VADER (old v1): negative — VADER -0.596 siebert+opinion (LIVE): negative — P(op)=1.00 P(pos)=0.00 Local NLI stance (deberta-v3-large): negative — NLI p=0.99 GPT-6.1 Sol: negative — Explicitly calls it a bad documentary and criticizes its missing context, unreliable narration, and careless approach. Claude Opus 5.5: negative — Calls it not a good documentary, criticizes Fielder as careless, and questions why people celebrate it. Claude Sonnet 5.5: negative — Commenter says it's not a good documentary, criticizes the narrator as careless and reckless, and doesn't understand why people celebrate it. Gemini 4 Argon: negative — The commenter explicitly states 'This isn't a good documentary' and criticizes the director's careless, reckless approach. Grok 4.7: negative — Explicitly calls it not a good documentary and criticizes the filmmaker's careless approach. GPT-6 Luna (cheap): negative — The commenter explicitly calls it a bad documentary and criticizes its careless, incomplete approach. Claude Haiku 5.5 (cheap): negative — VADER -0.596 Gemini 3.5 Flash-Lite (cheap): negative — The commenter explicitly states 'This isn't a good documentary' and criticizes the film's careless approach. | − | − | − | − | − | − | − | − | − | − | − |
| other-sub #1 w 0.354 So here are some quotes from the sort of reviews I’m talking about New York Times: Whether Holmes’s occasional blinks and tears in these chats are performative is open to question as are her grasp on language, truth, science and reality itself Slashfilm: Don't go into this… VADER (old v1): negative — VADER -0.81 siebert+opinion (LIVE): negative — P(op)=0.56 P(pos)=0.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.59 GPT-6.1 Sol: no_verdict — Criticizes reviewers’ interpretations of Holmes and ambiguity, without expressing a verdict on the film itself. Claude Opus 5.5: no_verdict — Criticizes critics' 'ambiguity' readings and Holmes's mystery, but gives no clear judgement of the film itself. Claude Sonnet 5.5: no_verdict — Mainly quotes other reviews; the commenter criticizes the 'ambiguity' reading of Holmes, not clearly judging the film itself. Gemini 4 Argon: no_verdict — The commenter criticizes the critical reception and framing of the subject matter without giving a verdict on the film itself. Grok 4.7: no_verdict — Quotes reviews and rejects ambiguity framing; no clear like or dislike of the film. GPT-6 Luna (cheap): no_verdict — They criticize reviewers’ interpretations of the film, but don’t state whether they liked or recommend the film itself. Claude Haiku 5.5 (cheap): negative — Author calls the reviews' ambiguity framing 'absolutely insane' and rejects the film's premise, signaling disapproval of its approach. Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter discusses various reviews and themes of the film without expressing a personal verdict. | − | − | 0 | 0 | 0 | 0 | 0 | 0 | 0 | − | 0 |
| other-sub #2 w 0.289 Light Spoilers for You Can See Everything ahead - be forewarned! It would be difficult to put onto the page all of the thoughts that have been crashing through my head since I saw “You Can See Everything” last night. But of course, the thing that has been keeping me up the most… VADER (old v1): positive — VADER 0.998 siebert+opinion (LIVE): neutral — P(op)=0.06 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.87 GPT-6.1 Sol: no_verdict — Discusses Holmes’s psychology and interprets a scene without expressing an overall judgment of the film. Claude Opus 5.5: no_verdict — Psychological analysis of Holmes and the film's thesis scene, with no explicit judgement of the film's quality. Claude Sonnet 5.5: no_verdict — Commenter analyzes Elizabeth Holmes's psychology and a key scene as thesis, but offers no explicit judgment of the film's quality; text is truncated. Gemini 4 Argon: no_verdict — The commenter analyzes the subject's psychology and a specific scene, without expressing whether they liked the film. Grok 4.7: no_verdict — Character and scene analysis without judging the film. GPT-6 Luna (cheap): no_verdict — The commenter analyzes the film’s themes and Elizabeth Holmes’s psychology without judging the film itself. Claude Haiku 5.5 (cheap): no_verdict — Psychological analysis of Elizabeth Holmes and the film's themes; no clear judgement of the film's quality or recommendation. Gemini 3.5 Flash-Lite (cheap): no_verdict — The commenter discusses the psychology of a character and film themes without expressing a clear verdict on the film. | + | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| other-sub #3 w 0.25 I know there’s a lot of discussion about how Holmes is portrayed in the film, and how uncomfortable the ambiguity is given what we know in real life. Which is of course true. Let’s be real, this person would understand who’s on first. But YCSE is not about whether Holmes is… VADER (old v1): positive — VADER 0.945 siebert+opinion (LIVE): neutral — P(op)=0.07 P(pos)=1.00 Local NLI stance (deberta-v3-large): mixed — NLI p=0.91 GPT-6.1 Sol: positive — Defends the film’s deliberate choices and praises its layered exploration of truth, belief, and contradiction. Claude Opus 5.5: positive — Admiring analysis of Fielder's layered structure and thematic depth; defends the film's choices as deliberate and effective. Claude Sonnet 5.5: positive — Commenter offers an admiring, in-depth analysis of Fielder's craft and layered structure, treating the film as thoughtful and effective. Gemini 4 Argon: positive — The commenter offers a thoughtful, appreciative analysis of the director's intent and the film's layered structure, calling it miraculous. Grok 4.7: positive — Defends and praises the film’s purposeful structure and thematic execution. GPT-6 Luna (cheap): positive — They praise the film’s layered structure and describe Holmes as a compelling vehicle for its themes. Claude Haiku 5.5 (cheap): positive — VADER 0.945 Gemini 3.5 Flash-Lite (cheap): no_verdict — The comment discusses the film's themes, structure, and characters analytically without expressing a personal verdict or recommendation. | + | 0 | 0 | + | + | + | + | + | + | + | 0 |