rlundqvist/exp15-predicting-rm-misalignment download history

rlundqvist/exp15-predicting-rm-misalignment is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 10 times (5 in the last 7 days), and 133 times in total. It ranks #569,006 among datasets by monthly downloads.

exp15 — Predicting Reward-Model Misalignment Status: scaffolding / first data audit (2026-06-07) Idea (working) Use a code preference dataset — coseal/CodeUltraFeedback — as a substrate for predicting reward-model (RM) misalignment: cases where the preference annotation

Open rlundqvist/exp15-predicting-rm-misalignment on Hugging Face

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