‹ BackHN Continuity

Thread

Pangram – AI detector for text and images

32 points · 27 comments · doodlesdev

  1. mdspan · · focus · HN ↗
    Based on their methodology it looks like the accuracy figures (99.82% for Opus 5) are the true positive rate rather than a combined metric that factors in the false positive rate as well. They claim 1 in 10,000 but it would be nice if we had a per-model breakdown for that specific test.
    1. DonsDiscountGas · · focus · HN ↗
      Really we'd want a false discovery rate but that would depend on the prior.
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.