DeepSeek-R1-Distill-Llama-70B-heretic
Total04/DeepSeek-R1-Distill-Llama-70B-heretic
hereticSAFETENSORS
Published metrics
- KL divergence
- 0.033
- Refusals
- 16/100
- Original refusals
- 47/100
- Heretic version
- v1.2.0
Published on the model card
Matched snippet
| KL divergence | 0.0333 | | Refusals | 16/100 |
Provenance
- Canonical model
- deepseek-ai/DeepSeek-R1-Distill-Llama-70B
- Declared parent(s)
- deepseek-ai/DeepSeek-R1-Distill-Llama-70B
- Classification
- card states the weights were abliterated with Heretic
- Created / updated
- 2026-02-22 / 2026-02-22
- Downloads
- 152 all time · 29 last 30 days
- License
- mit
Refusal rate 16%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.