Even (very) noisy LLM evaluators are useful for improving AI agents

 (tensorzero.com)

15 points | by GabrielBianconi 2 days ago

1 comments