PaulR11/training-runs-concerningness download history
PaulR11/training-runs-concerningness is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 340 times (299 in the last 7 days), and 4,129 times in total. It ranks #44,527 among datasets by monthly downloads.
GRPO Training Runs — Concerningness Scalar Reward (8-Target) Part of the data release for "Training Alignment Auditors via Reinforcement Learning" (ICLR 2026). Scalar-reward GRPO using the 38-metric Petri judge to compute a concerningness-based reward R = gate(realism) × ((c-3)/7)² with p