Souradip Chakraborty

Latest

• Post-training Large Language Models for Diverse High-Quality Responses
• Provably Sample Efficient RLHF via Active Preference Optimization