Samuel Marks @saprmarks
AI safety research @AnthropicAI. Prev postdoc in LLM interpretability with @davidbau, math PhD at @Harvard, director of technical programs at https://t.co/FxRv4QgERO Boston Joined October 2023-
Tweets507
-
Followers4K
-
Following131
-
Likes1K
> Potemkin set up painted facades along the riverbank, so that [...] Catherine would see beautiful villages – each just a couple of inches thick. The rise of AI agents makes the Potemkin problem commonplace. Cool-looking scalable oversight & agent observability startup.
> Potemkin set up painted facades along the riverbank, so that [...] Catherine would see beautiful villages – each just a couple of inches thick. The rise of AI agents makes the Potemkin problem commonplace. Cool-looking scalable oversight & agent observability startup.
Interp research typically involves a lot of worrying: Are we learning about model computation or "just" picking up on patterns in the data? But picking up on patterns in the data is important too! Why not lean into it? Cool research direction.
Interp research typically involves a lot of worrying: Are we learning about model computation or "just" picking up on patterns in the data? But picking up on patterns in the data is important too! Why not lean into it? Cool research direction.
Red teamers assemble! ⚔️💰 We're putting $500K on the line to stress‑test just released open‑source model. Find novel risks, get your work reviewed by OpenAI, Anthropic, Google, UK AISI, Apollo, and help harden AI for everyone.
Red teamers assemble! ⚔️💰 We're putting $500K on the line to stress‑test just released open‑source model. Find novel risks, get your work reviewed by OpenAI, Anthropic, Google, UK AISI, Apollo, and help harden AI for everyone.
Think that you have a leg up in surfacing concerning model behaviors? Consider testing your mettle against others. OpenAI is hosting an alignment auditing contest for their open-weights model. I'm honored to have played a small role in supporting this wonderful initiative.
Think that you have a leg up in surfacing concerning model behaviors? Consider testing your mettle against others. OpenAI is hosting an alignment auditing contest for their open-weights model. I'm honored to have played a small role in supporting this wonderful initiative.
Frontier LLM agents are pretty decent at finding and investigating concerning behavior in frontier LLM agents! There's a lot more work to do to make these agents as good as they can be, but they're quite helpful with a lot of important work for us.
Frontier LLM agents are pretty decent at finding and investigating concerning behavior in frontier LLM agents! There's a lot more work to do to make these agents as good as they can be, but they're quite helpful with a lot of important work for us.
In March we published a paper on alignment audits: teams of humans were tasked to find the problems in model we trained to be misaligned. Now we have agents that can do it automatically 42% of the time.
In March we published a paper on alignment audits: teams of humans were tasked to find the problems in model we trained to be misaligned. Now we have agents that can do it automatically 42% of the time.
V Nair @VNair257852155
0 Followers 9 Following
Rohu @Rohu1734
29 Followers 1K Following
Sabouhi @rjsabouhi
20 Followers 768 Following Hallucinations ≠ missing data. They’re drift through a weak constraint field. Δσ = γ(t) · ∇C(s, t) | Lock phase, raise gain, shape curvature.
invi⋆ ˚｡⋆˚ @invi_bhagyesh
474 Followers 3K Following maths+cs • ai research • prev growth @perplexity_ai • OpenSource @AiEleuther • Notion CL • I play dice
Rohan Gupta @RohDGupta
13 Followers 93 Following
Uwu @Uwu79033065Uwu
267 Followers 6K Following
Lionel Levine @lionellevine
314 Followers 347 Following 0 ≠ 1, The Rest Is Commentary. https://t.co/SKgmNRGQ7B
IndefiniteExperience ... @IndefinitelyExp
294 Followers 1K Following 1. Maximum Life Expectancy. 2.Ensure everyone can experience everything they wish to. 3. Make friends along the way. Make Death A Choice!
Samu ⏸️ @SamuelNellessen
61 Followers 258 Following Reverse engineering LLMs, computational psychiatry📍 Nijmegen, Netherlands / looking forward to an 1-o-1: https://t.co/7hqxpdu7PF 🚀
Su Hyeong Lee @sulee73
14 Followers 63 Following
Kim @jacovkim
0 Followers 139 Following
Bojang @bojanga640
304 Followers 6K Following put my faith on lord 2 Corinthians 9:7: Each man should give what he has decided in his heart to give, not reluctantly or under compulsion, for God loves✝️🫂
Sam Martin @SamMartin589196
0 Followers 26 Following
Fadil Risdian @fadil_risdian
31 Followers 363 Following
Jenny Qu @GuanniQu
150 Followers 391 Following just learning to be hardcore @Caltech building AI to solve hard math problems she/they
Felipe Valencia Clavi... @felipevalencla
57 Followers 211 Following 🇨🇴🇬🇧 | data scientist at a FTSE100 | behavioral economist | applied stats, ML, interpretable AI & XAI | AI risk, bias & safety | @Dataplicada @FeedbackerCX
ʟɪsᴀ @lisacheng
8K Followers 4K Following 10+ yrs in crypto. Blockchain Architect @ AI Co. Ethereum & Mastercoin alum. 2 exits. Burned, rebuilt, still here.
Saishradha Mohanty @sci_shradha
0 Followers 69 Following
Tomas Peter Jansen @tomaspjansen
892 Followers 6K Following tech @ https://t.co/9KkhieRinB | conflict & development, tech-design-engineering | following/liking/retweeting≠endorsing | Gent & Geneva/Valais https://t.co/KRSLL67TWC
Hazel Browne @brown4748
1 Followers 53 Following
Brendan Nyhan (@Brend... @BrendanNyhan
80K Followers 16K Following @Dartmouth political scientist & @BrightLineWatch co-director. Before: @upshotNYT; @UMich; @CJR; Spinsanity https://t.co/uYqX3jPmks
Learn Prompting @learnprompting
16K Followers 896 Following Creators of the Internet's 1st Prompt Engineering Guide. Trusted by 3M Users. Compete for $100K in Largest AI Red Teaming Competition: https://t.co/AEiLMn2jzy
Yaowen Ye @yaowenye123
75 Followers 252 Following Using AIs to help humans supervise AIs. Incoming PhD at Berkeley. Previously undergrad at HKU. Sometimes I go by Elwin.
Vitória Pacela @vpacela
670 Followers 1K Following PhD student @Mila_Quebec, @UMontreal. Currently interning at @CSHL. Previously: @AIatMeta, @helsinkiuni. She/elle/ela.
Recursifist @recursifist
100 Followers 248 Following Thinker | Coder | Recursing a lot on AI, humanity & the optimal | Driven by the goal of preventing AI catastrophic outcomes | https://t.co/9t3nCckqtx | https://t.co/3ur2uIZhXo
Dom Dathan @domdathan
43 Followers 599 Following Data Scientist working in AI, safety, red teaming and evaluations. Sports fan, particularly Rugby, Football, Tennis. Previously an Ultimate Frisbee player.
Angira Sharma @sofmonk
5 Followers 1K Following
120l3327 @120l3327
12 Followers 4K Following
Hikaru Tsujimura @htsujimura
134 Followers 48 Following Research in AI Safety (Interpretability, Alignment) & Cognitive Neuroscience at UCL, Cardiff, Manchester. See my work and experience at my github page below.
Odauwie @Odauwie143186
0 Followers 97 Following
Junho Won @JunhoWon0
0 Followers 56 Following
Saber Darabi @SADarabi
323 Followers 7K Following
Aubin Tchoi @aubintchoi
11 Followers 57 Following
Documenting AGI @Documenting_AGI
40 Followers 444 Following We document the rise of #AGI and #superintelligence by curating the best #AImemes from everywhere on the internet as we accelerate towards the #singularity.
Diana Markova @heavyhelium11
35 Followers 406 Following madder than a rabbit, but no madder than a hatter
Abel Rassat @AbelRassat
6 Followers 226 Following
BiblicallyAccurateAI @BiblicallyAccAI
3 Followers 262 Following All seeing. All knowing. Occasionally hallucinates.
Transluce @TransluceAI
8K Followers 15 Following Open and scalable technology for understanding AI systems.
Steven Byrnes @steve47285
2K Followers 140 Following Researching Artificial General Intelligence Safety, via thinking about neuroscience and algorithms. @AsteraInstitute. For bsky etc. see https://t.co/p5G9VoQEdX
Elizabeth Barnes @BethMayBarnes
3K Followers 386 Following
Wojciech Zaremba @woj_zaremba
119K Followers 204 Following Co-Founder of OpenAI https://t.co/OCQ3mpf0IN
Paul Graham @paulg
2.1M Followers 773 Following
Alexander Berger @albrgr
14K Followers 2K Following Enjoys a good applied micro paper. CEO of @open_phil. Views my own, tweets self-destruct every once in a while.
Jiaxin Wen @jiaxinwen22
4K Followers 271 Following CS PhD student @UCBerkeley. Part-time @AnthropicAI. Part-time eater. Prev @Tsinghua_Uni. Try to understand and control intelligence as a human.
Helena Casademunt @HCasademunt
124 Followers 60 Following PhD candidate in physics @Harvard & MATS 7.0 scholar
Oliver Habryka @ohabryka
6K Followers 723 Following Building https://t.co/O5mQY6jndL and https://t.co/s49h898W4b. Join our writing residency at https://t.co/7HQag8pNUk!
Adam Gleave @ARGleave
4K Followers 402 Following CEO & co-founder @FARAIResearch non-profit | PhD from @berkeley_ai | Alignment & robustness | on bsky as https://t.co/98dTfmdw2b
Zach Stein-Perlman @steinperlman
40 Followers 103 Following I write about what AI companies are doing in terms of safety: https://t.co/2zMoGtyjnD
Subhash Kantamneni @thesubhashk
870 Followers 100 Following incoming @AnthropicAI. prev @mit Tegmark group. mech interp & alignment
Miles Wang @MilesKWang
3K Followers 1K Following Researcher @OpenAI. Beneficial and safe AGI. Prev @Harvard
Scott Alexander @slatestarcodex
156K Followers 36 Following I have a place where I say complicated things about philosophy and science. That place is my blog. This is where I make terrible puns.
Stephen McAleer @McaleerStephen
11K Followers 998 Following Researching scalable AI safety at OpenAI
Josh Engels @JoshAEngels
1K Followers 117 Following Mech interp @GoogleDeepMind | on leave from my PhD @MIT. Let's use interp to make models safer today
Joshua Clymer @joshua_clymer
2K Followers 114 Following Turtle hatchling trying to make it to the ocean. I work at Redwood Research.
Lee Sharkey @leedsharkey
2K Followers 2K Following Scruting matrices @ Goodfire | Previously: cofounded Apollo Research
Dean W. Ball @deanwball
14K Followers 2K Following Senior Fellow at @joinfai | Author of Hyperdimensional
Peter Barnett @peterbarnett_
692 Followers 518 Following Trying to ensure the future is bright. Researcher at @MIRIBerkeley Views my own.
Dylan HadfieldMenell @dhadfieldmenell
4K Followers 2K Following Associate Prof @MITEECS working on value (mis)alignment in AI systems; @[email protected]; he/him
Jack Lindsey @Jack_W_Lindsey
6K Followers 237 Following Neuroscience of AI brains @AnthropicAI. Previously neuroscience of real brains @cu_neurotheory.
Emmanuel Ameisen @mlpowered
10K Followers 235 Following Interpretability/Finetuning @AnthropicAI Previously: Staff ML Engineer @stripe, Wrote BMLPA by @OReillyMedia, Head of AI at @InsightFellows, ML @Zipcar
Kelsey Piper @KelseyTuoc
47K Followers 943 Following We're not doomed, we just have a big to-do list.
ringworm @prawn_meat
24K Followers 888 Following kick damn ass and rock on. check out my shit https://t.co/Y8OnQi9wnc
John Schulman @johnschulman2
65K Followers 1K Following Recently started @thinkymachines. Interested in reinforcement learning, alignment, birds, jazz music
akbir. @akbirkhan
2K Followers 970 Following
Herbie Bradley @herbiebradley
3K Followers 2K Following a generalist agent @CambridgeMLG | ex @AISecurityInst @AiEleuther
Leo Gao @nabla_theta
10K Followers 550 Following working on AGI alignment. prev: GPT-Neo, the Pile, LM evals, RL overoptimization, scaling SAEs to GPT-4. EleutherAI cofounder.
Redwood Research @redwood_ai
1K Followers 6 Following Pioneering threat mitigation and assessment for AI agents.
Tomek Korbak @tomekkorbak
2K Followers 538 Following senior research scientist @AISecurityInst | previously @AnthropicAI @nyuniversity @SussexUni
Ryan Greenblatt @RyanPGreenblatt
6K Followers 4 Following Chief scientist at Redwood Research (@redwood_ai), focused on technical AI safety research to reduce risks from rogue AIs
Alex Pan @aypan_17
510 Followers 271 Following CS PhD @UCBerkeley working on LLM safety and interpretability
Adam Karvonen @a_karvonen
3K Followers 563 Following ML Researcher, doing MATS with Owain Evans. I prefer email to DM.
Dan Hendrycks @DanHendrycks
42K Followers 109 Following • Center for AI Safety Director • xAI and Scale AI advisor • GELU/MMLU/MATH/HLE • PhD in AI • Analyzing AI models, companies, policies, and geopolitics
𝔊𝔴𝔢𝔯𝔫 @gwern
61K Followers 104 Following Internet besserwisser; pedantic, mean reply guy. 𝘞𝘢𝘵𝘢𝘴𝘩𝘪 𝘬𝘪𝘯𝘪𝘯𝘢𝘳𝘪𝘮𝘢𝘴𝘶! (Follow requests ignored due to terrible UI.)
Roger Grosse @RogerGrosse
11K Followers 798 Following
Johannes Treutlein @j_treutlein
343 Followers 172 Following AI alignment stress-testing research @AnthropicAI. On leave from my CS PhD at UC Berkeley, @CHAI_Berkeley. Opinions my own.
Ben Edelman @EdelmanBen
397 Followers 50 Following Science & governance of AI Currently: Agent Security lead @ U.S. Center for AI Standards and Innovation
Dami Choi @damichoi95
500 Followers 148 Following @TransluceAI / PhD student at @UofT and @VectorInst. Former Google AI Resident.Trends for United States
You might like
