Idan Shenfeld

RL's Razor: Why Online Reinforcement Learning Forgets Less

Beyond Binary Rewards: Training LMs to Reason About Their Uncertainty

Best-of-N through the Smoothing Lens: KL Divergence and Regret Analysis

KL-Regularized RLHF with Multiple Reference Models: Exact Solutions and Sample Complexity

Learning How Hard to Think: Input-Adaptive Allocation of LM Computation

Language Model Personalization via Reward Factorization