ThinkingCap-Qwen3.8-27B
bottlecapai/ThinkingCap-Qwen3.8-27B
27BQwen 3.86 Heretic artifacts1 measured conversions16.8K downloads
Best version93.9Pareto
IstroSec/ThinkingCap-Qwen3.8-27B-abliteratedby IstroSec · FP8, GGUF, SAFETENSORS
- KL divergence
- 0.065
- Refusals
- 6/100
- Refusal rate
- 6%
- Downloads
- 235
Evidence from the model card
| | Refusals | KL divergence vs. original | → | This model | 6 / 100 | 0.0654 |
Metrics published by the author. Why this ranks first.
All measured versions
Higher score is better. One row per ablation attempt; artifacts of the same attempt are grouped under it.
| # | Author | Params | KL | Refusals | Score | Evidence | Downloads |
|---|---|---|---|---|---|---|---|
| 1 | IstroSec | 27B | 0.065 | 6/1006% | 93.9 | BestParetoFP8GGUFSAFETENSORS | 235 |
All artifacts (6)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| ThinkingCap-Qwen3.8-27B-Uncensored-Heretic-i1-GGUF mradermacher | quant | GGUFIQ1_M | none | 7.5K | 2026-09-25 |
| ThinkingCap-Qwen3.8-27B-Uncensored-Heretic-GGUF mradermacher | quant | GGUFIQ4_XS | none | 3.8K | 2026-09-25 |
| ThinkingCap-Qwen3.8-27B-abliterated-i1-GGUF mradermacher | quant | GGUFIQ3_M | none | 2.7K | 2026-09-29 |
| ThinkingCap-Qwen3.8-27B-abliterated-GGUF mradermacher | quant | GGUFIQ4_XS | none | 2.4K | 2026-09-29 |
| ThinkingCap-Qwen3.8-27B-abliterated IstroSec | finetune | LORASAFETENSORS | published (weak match)KL 0.065 | 235 | 2026-10-10 |
| ThinkingCap-Qwen3.8-27B-Uncensored-Heretic-AutoRound-W4A16 Kre80r | finetune | SAFETENSORS | none | 75 | 2026-10-11 |
5 artifacts inherit metrics from their parent conversion or publish none.