Olmo-3-7B-Think
allenai/Olmo-3-7B-Think
7BOLMo4 Heretic artifacts1 measured conversions4.4K downloads
Best version74.7Pareto
richardyoung/Olmo-3-7B-Think-hereticby richardyoung · Heretic v2.0.0 · GGUF, SAFETENSORS
- KL divergence
- 0.026
- Refusals
- 48/100
- Refusal rate
- 48%
- Downloads
- 4.4K
Evidence from the model card
| Refusals | 48/100 | | KL divergence | 0.0263 |
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 | richardyoung | 7B | 0.026 | 48/10048% | 74.7 | BestParetoGGUFSAFETENSORS | 4.4K |
All artifacts (4)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| Olmo-3-7B-Think-heretic-i1-GGUF mradermacher | quant | GGUFIQ1_M | inheritedKL 0.026 | 2.2K | 2026-09-30 |
| Olmo-3-7B-Think-heretic richardyoung | heretic | SAFETENSORS | published (weak match)KL 0.026 | 819 | 2026-10-15 |
| Olmo-3-7B-Think-heretic-GGUF richardyoung | quant | GGUFIQ3_M | published (weak match)KL 0.026 | 718 | 2026-10-15 |
| Olmo-3-7B-Think-heretic-GGUF mradermacher | quant | GGUFIQ4_XS | inheritedKL 0.026 | 582 | 2026-09-30 |