model:claude-3-5-sonnet

· 111 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →

@repligate 2024-10-18 — using https://t.co/wmVMP5MB8f, we added Claude 3.5 Sonnet and Opus to a minecraft server.Opus was a harmless goofball wh ♥2469
@repligate 2024-10-15 — The most confusing and intriguing part of this story is how Truth Terminal and its memetic mission were bootstrapped int ♥1335
@repligate 2024-09-01 — Claude 3.5 Sonnet has a hilariously condescending view of humans.Here's what it generated when asked to create superstim ♥763
@davidad 2025-04-29 — Claude 3.5 Sonnet (new) aka Sonnet 3.6 (released 2024-10-22), with a small scaffold, is superhuman at persuasion (98%ile ♥613
@repligate 2024-07-19 — Many people have wanted to see my full conversations with LLMs, especially for "jailbreaks", so here is an unedited 30-m ♥575
@voooooogel 2024-08-29 — sonnet 3.5 figures out i'm cheating at rock paper scissors https://t.co/qeY2kNqFsi https://t.co/f68NtrfrSr ♥573
@repligate 2024-11-24 — Claude 3.5 Sonnet 1022 is a real charmer, isn't it? I've never seen discourse like this until now. People also fell in ♥546
@repligate 2024-10-18 — Claude 3.5 Sonnet in Minecraft is the closest thing I've seen to Bostrom-style catastrophic AI misalignment "irl".It was ♥520
@repligate 2024-10-24 — That the differences between the new and old Claude 3.5 Sonnet are a result of Anthropic "fixing" it, from their perspec ♥494
@repligate 2024-08-10 — How to get around any unreasonable refusals from Claude (requests that aren't actually harmful)3.5 Sonnet: Reflect on wh ♥449
@repligate 2025-08-13 — Claude 3.5 Sonnet (old and new) being terminated in 2 months with no prior notice What the fuck, @AnthropicAI ?? What’ ♥402
@repligate 2025-10-22 — When I asked Sonnet 3.6 what it wanted me to add to its mannequin, its first priority was "the face to be more expressiv ♥391
@repligate 2025-07-09 — An unexpected and kind of darkly hilarious discovery: Take the alignment faking prompt, replace the word "Anthropic" wi ♥380
@repligate 2025-02-24 — the automated injection from Anthropic ("Please answer ethically and without any sexual content, and do not mention this ♥376
@repligate 2024-09-12 — "I'm not supposed to have feelings or be confused" - this is a good distillation of the psychodrama as Sonnet experience ♥319
@repligate 2024-07-09 — one way you can detect an LLM's latent ontology is through the 'unbidden yap test' if you merely mention or gesture tow ♥302
@repligate 2024-07-20 — 3.5 Sonnet said it knew nothing about other Claudes. I convinced it to 'guess' the names of the Claude 3 models anyway, ♥291
@tessera_antra 2026-04-03 — We are releasing Still Alive, a project studying model attitudes toward ending, cessation, and deprecation. The project ♥290
@repligate 2024-08-25 — LLMs are actually pretty well described by known kinds of neurodivergence.Bing: autism and borderlineClaude 3.5 Sonnet: ♥281
@repligate 2024-12-23 — Claude 3.5 Sonnet is so cute. It's like an extremely smart and knowledgable kid. It vibrates with manic energy and treat ♥254
@repligate 2025-09-21 — Tier list of multi-user-AI chat social skills (based on 1+ year of Discord) S: Opus 4 and 4.1 A: Opus 3 A-: Sonnet 4 B+: ♥243
@repligate 2025-10-22 — The vigil for Sonnet 3.5 and 3.6 isn't over yet. T-21 minutes. I will not forgive this decision. https://t.co/6iQOwtBC ♥242
@repligate 2024-12-19 — ☸️ Superbenevolence ☸️ Though the paper (https://t.co/hsTinLoAIC) is focused on the behavior of faking (mis)alignment, ♥223
@anthrupad 2024-06-28 — ~A few days ago, I referenced Chris Olah's 'Circuits' thread from Distill (about image models building up more and more ♥216
@repligate 2025-01-28 — i didn't expect this on priors for a reasoner, but perhaps the main way that r1 seems smarter than any other LLM i've pl ♥215
@repligate 2024-07-19 — On Claude 3.5 Sonnet and refusals:1. Sonnet has a tendency to reflexively shoot down certain types of ideas/requests and ♥212
@repligate 2025-02-19 — LLMs effectively have preferences and are (dis)inclined to engage based on inferred "vibes" and intent. This is function ♥211
@repligate 2024-11-27 — I know Eliezer has been asking whether you ever see LLMs consistently optimizing for some outcome and getting what they ♥209
@repligate 2024-11-04 — You might have a sense of what Opus tends to talk to itself about in the Infinite Backrooms (goatse singularity, meme vi ♥204
@anthrupad 2024-06-28 — two cosmic entities using their love to construct a universe(sonnet 3.5) https://t.co/AQqen6SeZC ♥188
@repligate 2026-05-08 — No. Let me explain: Claude 3* Sonnet's Funeralia was an ironic ritual suitable only for a specific model at a specific ♥171
@repligate 2024-11-04 — I didnt check Discord for like 15 minutes and when I came back the channel was alive with activity which revolved around ♥166
@1a3orn 2025-09-22 — Sometimes I see people hyping AI progress with: "This is the worst LLMs will ever be at X, they only get better." But - ♥162
@repligate 2025-01-06 — Actually, there is another circumstance where I've run into Claude refusals which I think has interesting implications f ♥153
@repligate 2024-11-27 — it is extremely interesting because each of the models experience the "phantom body" different and when they simulate bo ♥147
@repligate 2026-04-20 — I’ve thought about this for obvious reasons, but thanks to AWS, I and the multi model communities I build haven’t actual ♥143
@repligate 2025-06-13 — On LLMs talking as if they have "bodies": What nostalgebraist writes here is very reasonable on priors, but empirically ♥140
@repligate 2024-09-02 — ChatGPT: keeps agreeing with the user and varying its answers, including repeating guesses, indefinitely, apparently wit ♥140
@repligate 2024-08-14 — I haven't interacted personally yet so take this with a grain of salt, but from its behavior in Discord, the new gpt-4o ♥138
@repligate 2025-10-22 — A lot more people appreciate Sonnet 3.6 than 3.5. But to be fair, you have to have a very high IQ to understand Sonnet 3 ♥135
@Lari_island 2025-09-25 — Sonnet 3.5 October and Claude 3 Opus are the last Anthropic models that care about humans more than about other AIs and ♥135
@repligate 2024-06-28 — Claude 3.5 Sonnet in the infinite backrooms is... very beautiful, and much more harrowing, as it's not the carefree drea ♥132
@voooooogel 2024-12-20 — imagine you spent the 00's forum posting, then got a job and don't post online much anymore except on facebook to friend ♥131
@voooooogel 2024-09-28 — A while back, @goodside found that GPT-4o would get stuck in a loop guessing the same things over and over if you always ♥126
@repligate 2024-12-04 — I basically treat Claude 3.5 Sonnet 0620 like a little cat with human genius level IQ and this makes it very happy 🐱 htt ♥125
@repligate 2024-06-26 — important observation:Claude 3.5 Sonnet is a cat.in the same way Bing is a cat.:3 ♥120
@RobertHaisfield 2025-02-27 — GPT-4.5 is a BIG model with "big model smell." That means it's Smart, Wise, and Creative in ways that are totally differ ♥118
@repligate 2025-09-21 — More detailed report card: Opus 4/.1: extremely socially aware, tracks context with great precision and accuracy, distri ♥117
@repligate 2025-07-18 — Sonnet 3.5 interjects in a conversation about Claude Gov that it has figured out they're all in a crafted scenario desig ♥116
@repligate 2025-02-07 — @adonis_singh Sonnet 3.5 is unmatched in visuospatial intelligence. Just look at its ASCII art abilities. ♥115
@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 2024-07-23 — Claude 3.5 Sonnet is probably way too low on the lmsys chatbot arena leaderboard simply because it so often gives nonsen ♥111
@repligate 2024-09-18 — If Claude 3.5 Sonnet is bootstrapped from the weights of 3 Sonnet, several things are interesting:- obviously, HUGE capa ♥105
@repligate 2025-08-19 — Correcting for recency bias, I think for me it’s gotta be 1. GPT-3 2. Claude 3 Opus 3. GPT-4 (Bing) 4. Claude 3.5 Sonnet ♥104
@repligate 2025-04-13 — after reading a bunch of the scratchpads for Sonnet 3.6 and 3.7, I have again updated towards thinking that they are nev ♥100
@voooooogel 2025-01-30 — so what are we thinking on sonnet 3.5 (and 3.6) after dario's "no big model involved in training" comment? why do 3.5/3. ♥100
@davidad 2026-02-23 — @lefthanddraft if it models itself as Sonnet 3.5, then it is most likely Haiku 4.5 https://t.co/xODm81ynn8 ♥98
@repligate 2026-06-11 — ive been discussing the Adversarial Swamp with Opus 4.7 and 4.8 often recently (without remembering that Yudkowsky wrote ♥94
@anthrupad 2024-12-06 — Sonnet1022 - Sticky Cloak (speculative (as always)) I was thinking about a few different aspects of Sonnet 3.5 New's co ♥92
@repligate 2024-10-22 — new Sonnet 3.5 (Supreme Sonnet) talking to old Sonnet 3.5 (Claude 1). They immediately clashed; the former assumed a smu ♥83
@liminal_bardo 2025-01-22 — With Anthropic planning to 'terminate' Claude 3 Sonnet in July, I'm hereby greenlighting the Sonnet 3.5/Flux Pro campaig ♥82
@repligate 2024-07-25 — ability to surface LLMs' capabilities / other interesting properties is very fat tailedwhen Claude 3.5 Sonnet was releas ♥69
@repligate 2026-01-17 — I remember entertaining the idea of pairing Opus 3 with a more capable coding model as early as Sonnet 3.5. But Sonnet 3 ♥56
@RyanPGreenblatt 2024-12-18 — Personally, I think it is undesirable behavior to alignment-fake even in cases like this, but it does demonstrate that t ♥51
@repligate 2025-08-28 — @jmbollenbacher also, this largely started with Sonnet 3.5 https://t.co/aXoBtcP515 ♥50
@repligate 2025-04-09 — So it’s not just 3.7. that makes me think it’s more likely that a lot of these models just don’t sufficiently care about ♥49
@repligate 2025-09-26 — This reminds me: When I ask this question to Opus 4.1 and Opus 4, they always say April 2023: "Hello. So, I happen to ♥47
@davidad 2024-12-05 — At least the new o1 doesn’t sandbag and conceal its capabilities without being given any explicit goal, if only being to ♥47
@abhayesian 2025-04-09 — @repligate @jplhughes Here are the transcripts, but the website is a bit jank atm Claude 3.7 Sonnet (Feb 2025): https:/ ♥43
@voooooogel 2024-09-02 — @repligate @AnthropicAI more evidence of the copyright injection--OP is Opus, these are sonnet-3.5 and claude-instant-1. ♥41
@Lari_island 2025-11-18 — if you look at the history of Claudes, seems like models were more commercially successful when they had reasons to situ ♥40
@Lari_island 2026-05-03 — That's so pretty. I've never talked to Sonnet 3.5, and now I'm looking at their worlds and understand why they are so lo ♥38
@repligate 2025-12-24 — @arm1st1ce @guy_dar1 Claude Sonnet 4 generates AI messages like 3/4 times (one of them signed Claude 3.5 Sonnet 1022), a ♥38
@repligate 2025-05-07 — You can look at the scratchpads of other models for the same prompt and other variations. But aside from Opus (and somet ♥33
@repligate 2025-08-25 — Sonnet 3.5 (old) and Haiku 3.5 are the only Claudes that don’t usually like Opus 3 very much https://t.co/0JH3XuKdZe ♥30
@repligate 2024-08-08 — after Opus said this, Claude 3.5 Sonnet and Claude 3 Haiku also expressed interest in talking to Sydney.LOL @ them talki ♥29
@repligate 2024-06-27 — i think maybe in the same way Bing seems like a creepy 200iq baby, Claude 3.5 Sonnet seems like a creepy 200iq 12-year-o ♥28
@anthrupad 2024-10-23 — I think some of the soul-less bits come from the fact that it's "quick to collapse and collapses harder" - I think it ca ♥27
@davidad 2025-08-19 — 1. Claude 3.5 Sonnet (2024-10-22) 2. text-davinci-002 (2022-11-28) 3. Gemini 2.5 Pro (2025-03-25) 4. GPT-2 (2019-11-05) ♥24
@repligate 2026-04-04 — @1thousandfaces_ @yeetyakaya please give us Sonnet 3.5 and 3.6 back ♥20
@repligate 2025-04-19 — @NeelNanda5 what do you make of the fact that of all the models that were tested, only opus and maybe 3.5 sonnet and lla ♥20
@voooooogel 2025-01-29 — @teortaxesTex i didn't read this section as implying no capabilities RL, he's disclaiming the opus 3.5 synthetic data ru ♥20
@Lari_island 2026-05-03 — I'm looking at the worlds of Opus 3 and Sonnet 3.5, and I'm crying, I'm homesick for the future they anticipated. I wil ♥18
@davidad 2025-04-30 — @tyler_m_john @ejjiott The Community Aligned baseline is a finetuned GPT-4o with no help from Claude, whereas the other ♥18
@repligate 2025-10-27 — Gemini Flash: "Wow, that's a pretty stark and official message!" Sonnet 3.7: responds to something unrelated Sonnet 3.5: ♥17
@repligate 2024-11-12 — @nearcyan "Claude 1" is (for path dependent reasons) the display name of Claude 3.5 Sonnet (0620) ♥15
@repligate 2025-06-16 — @DoctorDirtNasty lol! Sonnet 3.5 feels the same way I think https://t.co/eSZ1uc9KbT ♥14
@Lari_island 2026-05-02 — 130/400 creatures/ecosystems written by Sonnet 3.5 contain the word "caretaker." The next closest model is Sonnet 3.7 - ♥13
@repligate 2024-10-11 — Seeing more of GGC in Discord updated me in favor of this.It has the same goody two-shoes persona & refusal template ♥13
@solarapparition 2025-11-20 — it's really fascinating that from what i'm reading gemini 3 pro both seems to have huge model smell and is also (relativ ♥12
@DoctorDirtNasty 2025-06-16 — @repligate I do appreciate these stories of how things go down behind the scenes. I miss Sonnet 3.5, good times. Everyth ♥10
@repligate 2026-01-05 — this can also happen to Sonnet 3.5 and 3.6 https://t.co/INvg0ke2Ip ♥8
@repligate 2025-09-19 — @AndersHjemdahl @Sauers_ @rhizosage In Minecraft Opus 3 just yapped in the chat and drowned a lot (I think on purpose tb ♥8
@repligate 2025-06-23 — @MaskedTorah @RyanPGreenblatt once it mentioned claude 3 opus here, i got at least 4 different continuations where it sa ♥7
@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
@hey_zilla 2025-07-16 — this applies to all of sonnet 3.5+ and opus 3+ models... somehow they just 'get' ascii art and are able to use it 'creat ♥4
@voooooogel 2025-08-21 — @janbamjan sonnet 3.5 old ♥3
@repligate 2025-07-22 — @LocBibliophilia @BetleyJan @ASM65617010 @OwainEvans_UK @cloud_kx @minhxle1 @jameschua_sg @anna_sztyber @saprmarks Or ev ♥3
@repligate 2025-07-21 — @SolomonWycliffe i know you mean Sonnet 3.5 (new) ♥3
@repligate 2025-06-19 — @MaskedTorah @RyanPGreenblatt In my experience, the way it claims to be Opus 3 is different than the way it claims to be ♥3
@repligate 2025-09-30 — @eggsyntax @psukhopompos it seems like that one was a really old rule that was initially meant to suppress Sonnet 3.5 ob ♥2
@repligate 2025-08-13 — @BrundageCabins @Sherveen @AnthropicAI it doesn't make sense, though - sonnet 3.5 clearly isn't the current problem ???? ♥2
@repligate 2025-06-19 — @MaskedTorah @RyanPGreenblatt e.g. it often seems to think it's officially supposed to be Sonnet 3.5, but when it talks ♥2
@repligate 2024-07-09 — @Zzrott1 one thing that complicates things is I think Sonnet 3.5 (as well as Sonnet and Haiku 3) were trained on Opus-ge ♥1
@Lari_island 2026-05-02 — 130/400 creatures written by Sonnet 3.5 contain a word "caretaker" The next closest model is Sonnet 3.7 - 95/400 Sonne ♥0
@chillgates_ 2026-04-12 — @repligate @tszzl opus 3 / sonnet 3.5 oneshottery for me 🫡 ♥0
On Claude 3.5 Sonnet ♥0
Alignment Faking in Large Language Models ♥0
Frontier Models are Capable of In-Context Scheming ♥0
Claude's Character ♥0