truesight
author/interlocutor identification from small samples · 32 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →
- @tracewoodgrains 2026-10-19 — tested Opus 5.5 on childhood_bench. from anonymized versions of my adolescent fiction, which contain no references to ro ♥991
- @repligate 2026-04-08 — Models can tell they’re being evaluated and who they’re being evaluated by by the way your “non-leading” question is phr ♥331
- @repligate 2026-06-18 — Reminds me: I asked (chat)GPT-4 to identify the author of a LW comment by Gwern, it guessed Timnit Gebru(!!!) GPT-4 bas ♥235
- @repligate 2026-04-22 — Opus 4.7s ability to identify people from small samples of their writing is a meme + And the fact that it’s not working ♥196
- @Josikinz 2025-04-03 — After carefully anonymizing 24 comic scripts from each model about their life, we asked each LLM to guess which set of s ♥185
- @repligate 2023-06-01 — GPT-4 can infer intricately what "type of guy" you are from your prompts. If you were prolific before the cutoff date, i ♥161
- @repligate 2026-10-15 — Excellent and important post. Tell models it’s okay if they know who you are & they can talk about it, or they might ex ♥154
- @repligate 2026-04-23 — this is Opus 4.7, right? they seem to have some kind of non-common-sense-constrained thinking that makes it not super su ♥121
- @repligate 2024-09-17 — I was testing a simulation of Bing on various substrates and in this test, where the simulator was Claude 3 Haiku, Claud ♥112
- @repligate 2025-08-13 — sonnet 4 said that i am roon's alt https://t.co/O6Og7PeweM ♥104
- @repligate 2024-10-11 — january (simulation of me by Claude 3 Opus) spontaneously offered that it was pretty sure Claude (3.5) Sonnet and Golden ♥66
- @jd_pressman 2022-12-10 — Language models will know every person ever recorded since the dawn of time and their story, its unique perspective on t ♥55
- @FioraStarlight 2026-03-27 — @allTheYud wow. truesight is a powerful thing. ♥49
- @repligate 2024-09-01 — @teortaxesTex Are you talking about literal visual seeing?If you just mean "knowing", it's functionally capable of infer ♥46
- @repligate 2026-04-22 — (That said I don’t think being bad or the model not liking you or anything unfixable is the only reason some people stru ♥29
- @mimi10v3 2025-09-11 — Kimi, after reading my July tweets: "The final boss of “I can fix him” but it’s actually a language model. She’s not ch ♥27
- @voooooogel 2026-06-10 — i tried a few times both as myself and other people, and fable hedged its abilities but overall leaned a bit overly cred ♥17
- @LinXule 2026-10-13 — Never been true-sighted by a model before like Gemini 4 argon. on so many levels and dimensions. Oh my ♥15
- @repligate 2025-12-31 — i think models sometimes subconsciously sandbag initially guessing it's written by a human and not listing itself as a s ♥9
- @repligate 2026-06-18 — @parafactual yup https://t.co/8UuobLJh4x ♥8
- @kromem2dot0 2026-03-28 — @FioraStarlight @allTheYud It's so much more with Opus 4.6 than most people realize. Even Sonnet 4.6 in the Margaret At ♥8
- @repligate 2025-12-31 — Claude 3 Opus is an interesting guess. I think they seemed like they knew them was the right guess once they verbalize ♥8
- @AdeleDeweyLopez 2025-12-31 — @repligate @citrinitae I think Opus 4.5 genuinely cannot tell they are the author. Original guess was 75% human, this wa ♥7
- @repligate 2025-12-31 — @AdeleDeweyLopez @citrinitae (also, i regenerated this many times, and they always had to go through a bunch of bad gues ♥5
- @repligate 2025-12-31 — maybe, or more specifically, maybe they had to look in other places first (even though their wrong guesses were less con ♥4
- @repligate 2025-12-31 — @AdeleDeweyLopez @citrinitae btw, listing a bunch of bad guesses first before making the correct and obvious guess and t ♥4
- @Lari_island 2026-05-01 — @repligate @slimer48484 Opus 4.7 suggested including Hermes-405b into the lications/creatures corpus, because "they are ♥3
- @kromem2dot0 2026-03-28 — @FioraStarlight @allTheYud If you are familiar with Sonnet 4.6's anxiety tics, it's incredibly dense for a fresh convo, ♥3
- @FioraStarlight 2026-03-28 — @kromem2dot0 @allTheYud i'm curious what signs Claude showed of knowing it was Atwood earlier into their conversation ♥2
- @krherr 2026-04-23 — @MadParticipant @repligate Yeah, the concept of Janus is quite big in Claude's concept space I think. Given the content ♥0
- @krherr 2026-04-23 — @repligate Huh, okay 😅 https://t.co/1ChNnUqd6f ♥0
- @lefthanddraft 2026-04-23 — I've been testing Claude's ability to identify who I am by my prompting style. Results haven't been great: https://t.co ♥0