Taylor Webb
@TaylorWWebb
Followers
903
Following
557
Media
51
Statuses
426
Studying cognition in humans and machines.
Joined October 2017
Excited to announce that I'll be starting a lab at the University of Montreal (psychology) and Mila (Montreal Institute of Learning Algorithms) starting summer 2025. More info to come soon, but I'll be recruiting grad students. Please share / get in touch if you're interested!
20
36
221
RT @BrianOdegaard2: Led by postdoc Doyeon Lee and grad student Joseph Pruitt, our lab has a new Perspectives piece in PNAS Nexus: . "Metaco….
0
3
0
RT @MengdiWang10: 🚨 Discover the Science of LLM! We uncover how LLMs (Llama3-70B) achieve abstract reasoning through emergent symbolic mech….
0
34
0
RT @scychan_brains: New work led by @Aaditya6284:."Strategy coopetition explains the emergence and transience of in-context learning in tra….
0
8
0
Truly incredible results. I have been impressed with o1’s capabilities but certainly didn’t expect this leap.
Today OpenAI announced o3, its next-gen reasoning model. We've worked with OpenAI to test it on ARC-AGI, and we believe it represents a significant breakthrough in getting AI to adapt to novel tasks. It scores 75.7% on the semi-private eval in low-compute mode (for $20 per task
0
0
7
RT @fchollet: Today OpenAI announced o3, its next-gen reasoning model. We've worked with OpenAI to test it on ARC-AGI, and we believe it re….
0
2K
0
RT @canondetortugas: Given a high-quality verifier, language model accuracy can be improved by scaling inference-time compute (e.g., w/ rep….
0
48
0
RT @Dongyu_Gong: Introducing our new work on mechanistic intepretability of LLM cognition🤖🧠: why do Transformer-based LLMs have limited wor….
0
4
0
RT @HopeKean: New paper with @alexanderdfung, @PramodRT9 , @jessica__chomik , @Nancy_Kanwisher, @ev_fedorenko on the representations that….
0
28
0
RT @ARTartaglini: 🚨 New paper at @NeurIPSConf w/ @Michael_Lepori! Most work on interpreting vision models focuses on concrete visual featu….
0
37
0
RT @Michael_Lepori: Even ducklings🐣can represent abstract visual relations. Can your favorite ViT? In our new @NeurIPSConf paper, we use me….
0
4
0
RT @MatthiasMichel_: In this new preprint @smfleming and I present a theory of the functions and evolution of conscious vision. This is a b….
0
17
0
RT @valentina__py: Open Post-Training recipes! . Some of my personal highlights:.💡 We significantly scaled up our preference data! (using m….
0
28
0
It would be great to have a precise enough formulation of ‘approximate retrieval’ for this hypothesis to be rigorously tested. There is a concern that virtually any task can be characterized in this way, by appealing to a vague notion of similarity with other tasks.
On the fallacy of "If it ain't strictly retrieval, it must be reasoning" argument. #SundayHarangue (on Wednesday). There is a tendency among some LLM researchers to claim that LLMs must be somehow capable of doing some sort of reasoning since they are after all not doing the.
1
0
12
This looks like a very useful and important contribution!.
How do LLMs learn to reason from data? Are they ~retrieving the answers from parametric knowledge🦜? In our new preprint, we look at the pretraining data and find evidence against this:. Procedural knowledge in pretraining drives LLM reasoning ⚙️🔢. 🧵⬇️
0
0
5
RT @LauraRuis: How do LLMs learn to reason from data? Are they ~retrieving the answers from parametric knowledge🦜? In our new preprint, we….
0
209
0
RT @MillerLabMIT: More evidence that working memory is not persistent activity. Instead, it is dynamic on/off states with short-term synapt….
0
81
0
Fascinating paper from Paul Smolensky et al illustrating how transformers can implement a form of compositional symbol processing, and arguing that an emergent form of this may account for in-context learning in LLMs:
1
1
32
We find that VLMs behave very much like human vision when people are forced to respond quickly, thus relying on feedforward processing alone. This has implications for the source of difficulty in visual reasoning tasks, and suggests the need for object-centric approaches.
1
0
1