Show HN: Verdict – model evals on your own data, not someone else's benchmarkgithub.com·2 pts·agunapal·0