uncensor-v1-dpo
- Type
- dataset
- Venue
- rx-dev
- Year
- 2026
- Source
- huggingface
- Access
- free
- Added
- 2026-07-17T20:01:00.537719+00:00
- Verified
- 2026-07-17T20:01:00.537719+00:00
Summary
This is literally just LLM-LAT/harmful-dataset but flipped, so the rejected answer is the refusal, and the accepted answer is the original incorrect one
Keywords
hf-dataset json text datasets pandas mlcroissant polars
Research notes
- downloads=14; likes=16