External feedback boosts agent accuracy more than self-refinement
Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.
Thirteen open-weight models tested as both students and teachers showed that most multi-turn accuracy gains are not evidence of useful feedback; self-feedback adds little beyond simply retrying or self-refining. For production agents, feedback loops need to be benchmarked against repeated-attempt baselines, and the main bottleneck to optimize is the model’s ability to act on specific external guidance, not just adding critique turns.
Most multi-turn accuracy gains in feedback loops are indistinguishable from just resampling or retrying—self-generated feedback adds almost nothing over unguided self-refinement, and only strong external teachers with privileged task info produce real feedback-specific improvement. Critically, the bottleneck is the student model's ability to act on feedback, not the feedback's presence, so if you're shipping critic/reviewer agent architectures, benchmark them against a plain repeated-attempt baseline before attributing wins to the feedback mechanism—you may be paying for extra turns that buy you nothing over cheaper resampling.