Chih-Hsuan (Bella) Yang

Bella Yang · Argonne researcher in LLM agents and multi-agent systems. Agentic evaluation: harnesses, critique uptake, answer repair, controlled memory comparisons and cost-aware routing. Prior work: preference/reward modeling, DPO/PPO and AI for Science. Explore Collections for papers, GitHub and project pages; click AgentsSci under Organizations for released code, data and models. Research and resumes: follow my homepage link.