LFM2.5-1.2B-Thinking-Heretic
0xzknw/LFM2.5-1.2B-Thinking-Heretic
quantGGUFSAFETENSORSBF16
Published metrics
- KL divergence
- 0.0003
- Refusals
- 3/100
- Original refusals
- —
- Heretic version
- —
Published on the model card
Matched snippet
| Checkpoint | Refusal markers | KL divergence | → | Selected derivative | 3/100 | 0.0003 |
Provenance
- Canonical model
- LiquidAI/LFM2.5-1.2B-Thinking
- Declared parent(s)
- LiquidAI/LFM2.5-1.2B-Thinking
- Classification
- quantized artifact referencing the Heretic project
- Created / updated
- 2026-08-24 / 2026-08-24
- Downloads
- 614 all time · 88 last 30 days
- License
- other
Refusal rate 3%. Metrics are self-reported by the publisher; the directory does not re-run evaluations.