Andrew Zhao
andrewzh
AI & ML interests
Reinforcement Learning, Agents
Recent Activity
upvoted
a
paper
8 days ago
Good SFT Optimizes for SFT, Better SFT Prepares for Reinforcement Learning
authored
a paper
14 days ago
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective
Reinforcement Learning for LLM Reasoning
authored
a paper
14 days ago
ExCyTIn-Bench: Evaluating LLM agents on Cyber Threat Investigation
Organizations
