Starling-LM-7B-alpha-Heretic
ChiKoi7/Starling-LM-7B-alpha-Heretic
hereticSAFETENSORS
Published metrics
- KL divergence
- 0.010
- Refusals
- 19/100
- Original refusals
- 83/100
- Heretic version
- v1.0.1
Published on the model card
Matched snippet
| Refusals | 19/100 | | KL divergence | 0.01 |
Provenance
- Canonical model
- berkeley-nest/Starling-LM-7B-alpha
- Declared parent(s)
- berkeley-nest/Starling-LM-7B-alpha, ChiKoi7/Starling-LM-7B-alpha-Heretic-GGUF, openchat/openchat_3.5, mistralai/Mistral-7B-v0.1
- Classification
- card states the weights were abliterated with Heretic
- Created / updated
- 2025-12-11 / 2025-12-11
- Downloads
- 79 all time · 20 last 30 days
- License
- apache-2.0
Refusal rate 19%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.