←
Hacker News
Even (very) noisy LLM evaluators are useful for improving AI agents
11 points
0 comments
2 days ago