AI.info

Peter Chen

Explore Peter Chen on AI.info.

• A Zeroth-Order Paradigm for LLM Preference Alignment
• Reward-free Alignment for Conflicting Objectives
• GenEnv: Difficulty-Aligned Co-Evolution Between LLM Agents and Environment Simulators
• Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Reward