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

11 points by GabrielBianconi 2 days ago

0 comments:

Data from: Hacker News, provided by Hacker News (unofficial) API