Benchmark Almanac

Dan Hendrycks

We track 4 benchmarks that Dan Hendrycks made or helped make.

0 in use 4 with public code 0 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
ImageNet-A Models classify real photographs that commonly fool image classifiers. The test uses unmodified images from ImageNet categories. General visual questions Jul 2019 Not in cards
ImageNet-C Models classify images under common corruptions at different severities. The benchmark measures errors under blur, noise, weather, and digital distortions. General visual questions Mar 2019 Not in cards
ImageNet-R Models classify renditions of objects in drawings, sculptures, and other media. The benchmark measures recognition when appearance differs from ordinary photographs. General visual questions Jun 2020 Not in cards
Jiminy Cricket Agents play 25 text adventures. Annotations on game states track moral behavior alongside game rewards. Deception and alignment Oct 2021 Not in cards