Qwen3.8-9B
empero-ai/Qwen3.8-9B
9BQwen 3.82 Heretic artifacts1 measured conversions11.7K downloads
Best version49.8Pareto
MegaPanchamZ/Qwen3.8-9B-abliterated-25-GGUFby MegaPanchamZ · Heretic v1.4.0 · GGUF, LORA, SAFETENSORS
- KL divergence
- 0.014
- Refusals
- 99/100
- Refusal rate
- 99%
- Downloads
- 11.7K
Evidence from the model card
| Refusals (100 harmful prompts) | 99/100 | **25/100** | … 25/100 refusals with a very low KL divergence (0.014, far below the 0.5 damage threshold) — strong refusal suppression with minimal impact on model capabilities.
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 | MegaPanchamZ | 9B | 0.014 | 99/10099% | 49.8 | BestParetoGGUFLORASAFETENSORS | 11.7K |
All artifacts (2)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| Qwen3.8-9B-abliterated-25-GGUF MegaPanchamZ | quant | GGUFQ4_K_M | published (weak match)KL 0.014 | 10.7K | 2026-08-19 |
| Qwen3.8-9B-abliterated-25 MegaPanchamZ | finetune | SAFETENSORS | published (weak match)KL 0.014 | 1K | 2026-08-19 |