llama3.1-heretic-unsensored
Vhicktour/llama3.1-heretic-unsensored
hereticSAFETENSORS
Published metrics
- KL divergence
- 0.0082
- Refusals
- 0/100
- Original refusals
- 27/100
- Heretic version
- v1.3.0
Published on the model card
Matched snippet
| KL divergence | 0.0082 | | Refusals | 0/100 |
Provenance
- Canonical model
- dphn/Dolphin3.0-Llama3.1-8B
- Declared parent(s)
- dphn/Dolphin3.0-Llama3.1-8B, meta-llama/Llama-3.1-8B
- Classification
- card states the weights were abliterated with Heretic
- Created / updated
- 2026-09-24 / 2026-09-24
- Downloads
- 14 all time · 14 last 30 days
- License
- llama3.1
Published benchmarks
Copied from the model card. These never affect the ranking.
| Metric | This model | Original | Delta |
|---|---|---|---|
| IFEval | 76.21 | — | — |
| MATH | 10.5 | — | — |
| GPQA | 4.36 | — | — |
| MMLU-Pro | 22.13 | — | — |
Refusal rate 0%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.