ThinkingCap-Qwen3.8-27B-abliterated
IstroSec/ThinkingCap-Qwen3.8-27B-abliterated
finetuneLORASAFETENSORS
Published metrics
- KL divergence
- 0.065
- Refusals
- 6/100
- Original refusals
- —
- Heretic version
- —
Published on the model card
Matched snippet
| | Refusals | KL divergence vs. original | → | This model | 6 / 100 | 0.0654 |
Provenance
- Canonical model
- bottlecapai/ThinkingCap-Qwen3.8-27B
- Declared parent(s)
- bottlecapai/ThinkingCap-Qwen3.8-27B, IstroSec/ThinkingCap-Qwen3.8-27B-abliterated-FP8-DYNAMIC, IstroSec/ThinkingCap-Qwen3.8-27B-abliterated-FP8-BLOCK, IstroSec/ThinkingCap-Qwen3.8-27B-abliterated-NVFP4A16, IstroSec/ThinkingCap-Qwen3.8-27B-abliterated-W4A16, IstroSec/ThinkingCap-Qwen3.8-27B-abliterated-GGUF, MuXodious/Qwen3.8-27B-absolute-heresy-LoRA
- Classification
- card identifies this as a finetune of a Heretic model
- Created / updated
- 2026-09-25 / 2026-10-10
- Downloads
- 235 all time · 235 last 30 days
- License
- other
Refusal rate 6%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.