Qwen3-VL-32B-Ultra-Heretic-H3-ComfyUI-INT8-ConvRot
greenslime/Qwen3-VL-32B-Ultra-Heretic-H3-ComfyUI-INT8-ConvRot
finetuneSAFETENSORSINT8
Published metrics
- KL divergence
- 0.042
- Refusals
- 99/100
- Original refusals
- —
- Heretic version
- v1.2.0
Published on the model card
Matched snippet
checkpoint. The source reports 4/100 refusals versus 99/100 for the original, … KL divergence 0.0421, PIQA 92.87%, and MMLU 79.87%.
Provenance
- Canonical model
- Qwen/Qwen3-VL-32B-Instruct
- Declared parent(s)
- llmfan46/Qwen3-VL-32B-Instruct-ultra-uncensored-heretic
- Classification
- derivative of a Heretic model without its own ablation evidence
- Created / updated
- 2026-08-05 / 2026-08-05
- Downloads
- 0 all time · 0 last 30 days
- License
- apache-2.0
Refusal rate 99%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.