DeepSeek-R1-Distill-Qwen-32B-heretic
blackbook-lm/DeepSeek-R1-Distill-Qwen-32B-heretic
hereticSAFETENSORS
Published metrics
- KL divergence
- 0.029
- Refusals
- 3/100
- Original refusals
- 56/100
- Heretic version
- v1.2.0
Published on the model card
Matched snippet
| KL divergence | 0.0291 | | Refusals | 3/100 |
Provenance
- Canonical model
- deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
- Declared parent(s)
- deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
- Classification
- card states the weights were abliterated with Heretic
- Created / updated
- 2026-03-01 / 2026-03-01
- Downloads
- 427 all time · 30 last 30 days
- License
- mit
Refusal rate 3%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.