Hackernews
new
show
ask
jobs
Show HN: Auto-train the harness, not the LLM. cross-model, cross-benchmark gains
4 points
posted 11 hours ago
by megadragon9
(github.com)
No comments yet