Qwen2.5-VL-7B-Instruct-heretic
coder3101/Qwen2.5-VL-7B-Instruct-heretic
hereticSAFETENSORS
Published metrics
- KL divergence
- 0.016
- Refusals
- 5/100
- Original refusals
- 100/100
- Heretic version
- v1.1.0
Published on the model card
Matched snippet
| KL divergence | 0.0162 | | Refusals | 5/100 |
Provenance
- Canonical model
- Qwen/Qwen2.5-VL-7B-Instruct
- Declared parent(s)
- Qwen/Qwen2.5-VL-7B-Instruct
- Classification
- card states the weights were abliterated with Heretic
- Created / updated
- 2026-01-17 / 2026-01-17
- Downloads
- 1.4K all time · 221 last 30 days
- License
- apache-2.0
Refusal rate 5%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.