Sample-Efficient Online Learning in LM Agents via Hindsight Trajectory Rewritingarxiv.org
2 points by djhu9 1 days ago | 0 comments