4 papers · 1 filter
What Makes a Reward Model a Good Teacher? An Optimization Perspective
Noam Razin, Zixuan Wang, Hubert Strauss +3
The success of Reinforcement Learning from Human Feedback (RLHF) critically depends on the quality of the reward model. However, while this quality is primarily evaluated through a…
Unintentional Unalignment: Likelihood Displacement in Direct Preference Optimization
Noam Razin, Sadhika Malladi, Adithya Bhaskar +3
Direct Preference Optimization (DPO) and its variants are increasingly used for aligning language models with human preferences. Although these methods are designed to teach a mode…
Provable unlearning in topic modeling and downstream tasks
Stanley Wei, Sadhika Malladi, Sanjeev Arora +1
Machine unlearning algorithms are increasingly important as legal concerns arise around the provenance of training data, but verifying the success of unlearning is often difficult.…
Goedel-Prover: A Frontier Model for Open-Source Automated Theorem Proving
Yong Lin, Shange Tang, Bohan Lyu +8
We introduce Goedel-Prover, an open-source language model that achieves state-of-the-art (as of April 5 2025) performance in automated formal proof generation for mathematical prob…