Shockingly Simple Self-retrospection Improves Agentic Models Without RL
People learn not only by repeating successful actions, but also by recounting and explaining their experiences, revising their understanding to guide future behavior. Can a language-model agent improve its future actions by training only on explanations of its own experience? We investigate this question by studying Retrospection-Only Fine-Tuning (ROFT), a minimal online procedure designed to…
We haven't written up this one. arXiv cs.AI has the full story — the link below goes straight to it.
