LFM2.5-24B-A2B-Uncensored-Heretic-GGUF
OS-Software/LFM2.5-24B-A2B-Uncensored-Heretic-GGUF
quantGGUFQ4_K_M
Published metrics
- KL divergence
- 0.017
- Refusals
- 99/100
- Original refusals
- —
- Heretic version
- —
Published on the model card
Matched snippet
| File | Refusals (original → this) | KL divergence | MMLU vs original | → | QAD Q40 | 99/100 → 0/100 | 0.0174 | −0.62 (all 14,042, p = 0.004) |
Provenance
- Canonical model
- LiquidAI/LFM2.5-24B-A2B
- Declared parent(s)
- LiquidAI/LFM2.5-24B-A2B, OS-Software/LFM2.5-24B-A2B-Uncensored-Heretic
- Classification
- quantized artifact of a Heretic model
- Created / updated
- 2026-10-16 / 2026-10-16
- Downloads
- 403 all time · 403 last 30 days
- License
- other
Published benchmarks
Copied from the model card. These never affect the ranking.
| Metric | This model | Original | Delta |
|---|---|---|---|
| MMLU | 55.35 | 55.97 | -0.6 |
| HumanEval | 87.2 | 88.41 | -1.2 |
Refusal rate 99%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.