I'll be presenting my work on factual knowledge acquisition at #NeurIPS2024! 📅 Main Poster Session 2 East, Dec 11th 16:30-19:30

🎉 Excited to share that our paper "How Does Vision-Language Adaptation Impact the Safety of Vision Language Models?" has been accepted to #ICLR2025! 🖼 Vision-Language Adaptation empowers LLMs to process visual information—but how does it impact their safety? 🛡 And what about

𝗖𝗮𝘂𝘀𝗮𝘁𝗶𝗼𝗻, 𝗺𝗮𝗻𝘆 𝘁𝗵𝗶𝗻𝗴𝘀 Always thought causation connected very closely to the notion of modularity in general but specifically in neuroscience. Finally found some discussion by Cartwright!

After 7 years, thrilled to finally share our #MICrONS functional connectomics results! We recorded activity from ~75K neurons in visual cortex in a single mouse, then mapped its wiring using electron microscopy. To systematically characterize neuron function, we built the first

🙁 LLMs are overconfident even when they are dead wrong. 🧐 What about reasoning models? Can they actually tell us “My answer is only 60% likely to be correct”? ❗Our paper suggests that they can! Through extensive analysis, we investigate what enables this emergent ability.

🚨 New Paper co-led with byeongguk jeon 🚨 Q. Can we adapt Language Models, trained to predict next token, to reason in sentence-level? I think LMs operating in higher-level abstraction would be a promising path towards advancing its reasoning, and I am excited to share our

[1/6] Ever wondered why Direct Preference Optimization is so effective for aligning LLMs? 🤔 Our new paper dives deep into the theory behind DPO's success, through the lens of information gain. Paper: "Differential Information: An Information-Theoretic Perspective on Preference

Transformer-based neural networks achieve impressive performance on coding, math & reasoning tasks that require keeping track of variables and their values. But how can they do that without explicit memory? 📄 Our new ICML paper investigates this in a synthetic setting! 🧵 1/13

🚨 New Paper 🧵 How effectively do reasoning models reevaluate their thought? We find that: - Models excel at identifying unhelpful thoughts but struggle to recover from them - Smaller models can be more robust - Self-reevaluation ability is far from true meta-cognitive awareness