PaulR11/scaling-crux-eval download history

PaulR11/scaling-crux-eval is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 293 times (41 in the last 7 days), and 915 times in total. It ranks #50,050 among datasets by monthly downloads.

Scaling Crux Evaluation (25 models × 50 cruxes) Part of the data release for "Training Alignment Auditors via Reinforcement Learning" (ICLR 2026). Large-scale crux-eval sweep measuring V1/V2/pairwise rates at fixed crux continuation points across 25 frontier + trained models. Used in crux

Open PaulR11/scaling-crux-eval on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.