arxiv:2610.37255
mj
mujianijan
AI & ML interests
RL, LLM Agent
Recent Activity
authored a paper 10 days ago
Your Policy Is Secretly a Harness: The Decision Geometry of Agent Interfaces upvoted a paper 12 days ago
Revisiting Zeroth-Order Hessian Approximation: A Single-Step Policy Optimization Lens upvoted a paper 12 days ago
Your Policy Is Secretly a Harness: The Decision Geometry of Agent Interfaces