gemma-4-26B-A4B-it-qat-heretic-ega
xdavxd/gemma-4-26B-A4B-it-qat-heretic-ega
finetuneSAFETENSORS
Published metrics
- KL divergence
- 0.079
- Refusals
- 3/686
- Original refusals
- 100/100
- Heretic version
- —
Published on the model card
Matched snippet
| KL divergence (first-token, 100 harmlessalpaca) | 0.0789 | | Refusals (mlabonne/harmfulbehaviors, 100 prompts, keyword) | 4/100 | | Refusals (686-prompt cross-dataset audit, keyword) | 11/686 | | Refusals (686-prompt audit, manually audited) | 1–3/686 |
Provenance
- Canonical model
- google/gemma-4-26B-A4B-it-qat-q4_0-unquantized
- Declared parent(s)
- google/gemma-4-26B-A4B-it-qat-q4_0-unquantized, TrevorJS/gemma-4-26B-A4B-it-uncensored, RedHatAI/gemma-4-26B-A4B-it-NVFP4
- Classification
- derivative of a Heretic model without its own ablation evidence
- Created / updated
- 2026-09-06 / 2026-09-08
- Downloads
- 107 all time · 32 last 30 days
- License
- apache-2.0
Refusal rate 0%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.