Social Post Explorers

AI & ML interests

None defined yet.

Recent Activity

wenbopan  authored a paper 6 days ago

Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts

wenbopan  authored a paper 9 days ago

The Hidden Dimensions of LLM Alignment: A Multi-Dimensional Analysis of Orthogonal Safety Directions

wenbopan  submitted a paper 12 days ago

Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts

View all activity

Team members 850