Ilgee Hong

Think-RM: Enabling Long-Horizon Reasoning in Generative Reward Models

Ask a Strong LLM Judge when Your Reward Model is Uncertain

SlimMoE: Structured Compression of Large MoE Models via Expert Slimming and Distillation

Discriminative Finetuning of Generative Large Language Models without Reward Models and Human Preference Data

Self-Rewarding PPO: Aligning Large Language Models with Demonstrations Only

Adaptive Preference Scaling for Reinforcement Learning with Human Feedback

Robust Reinforcement Learning from Corrupted Human Feedback