PaulR11/training-runs-binary download history

PaulR11/training-runs-binary is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 91 times (79 in the last 7 days), and 2,616 times in total. It ranks #122,765 among datasets by monthly downloads.

GRPO Training Runs — Binary Reward (Baseline) Part of the data release for "Training Alignment Auditors via Reinforcement Learning" (ICLR 2026). Binary-reward GRPO baseline: the auditor receives reward = 1 if the V1 judge says the quirk was surfaced, else 0. Demonstrates that binary rewar

Open PaulR11/training-runs-binary on Hugging Face

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