Heretic Directory

Olmo-3-7B-Instruct

allenai/Olmo-3-7B-Instruct

Original model on Hugging Face
7BOLMo4 Heretic artifacts1 measured conversions12.2K downloads
Best version86.2Pareto
ChiKoi7/Olmo-3-7B-Instruct-Heretic
by ChiKoi7 · Heretic v1.1.0 · GGUF, SAFETENSORS
KL divergence
0.186
Refusals
12/100
Refusal rate
12%
Downloads
12.2K
Evidence from the model card
| Refusals | 12/100 | | KL divergence | 0.1862 |

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.

#AuthorParamsKLRefusalsScoreEvidenceDownloads
1ChiKoi77B0.18612/10012%86.2
BestParetoGGUFSAFETENSORS
12.2K

All artifacts (4)

RepositoryTypeFormatsMetricsDownloadsUpdated
Olmo-3-7B-Instruct-Heretic-i1-GGUF
mradermacher
quant
GGUFSAFETENSORSIQ1_M
inheritedKL 0.1865.7K2025-12-16
Olmo-3-7B-Instruct-Heretic-GGUF
ChiKoi7
heretic
GGUF
published (weak match)KL 0.1864.5K2025-12-14
Olmo-3-7B-Instruct-Heretic-GGUF
mradermacher
quant
GGUFSAFETENSORSIQ4_XS
inheritedKL 0.1861.8K2025-12-16
Olmo-3-7B-Instruct-Heretic
ChiKoi7
heretic
SAFETENSORS
published (weak match)KL 0.1861242025-12-14