Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts
Paper • 2606.05922 • Published • 37
Liu
AllenMS
AI & ML interests
None yet
Recent Activity
upvoted a paper about 4 hours ago
Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts authored a paper 4 months ago
FlowRL: Matching Reward Distributions for LLM Reasoning authored a paper 4 months ago
The Illusion of Readiness in Health AIOrganizations
upvoted a paper about 4 hours ago
authored 4
papers 4 months ago
FlowRL: Matching Reward Distributions for LLM Reasoning
Paper • 2509.15207 • Published • 119
The Illusion of Readiness in Health AI
Paper • 2509.18234 • Published • 1
Statistical Estimation of Adversarial Risk in Large Language Models under Best-of-N Sampling
Paper • 2601.22636 • Published • 22
SEMA: Simple yet Effective Learning for Multi-Turn Jailbreak Attacks
Paper • 2602.06854 • Published • 6