Agents-A1-4B
InternScience/Agents-A1-4B
4BAgents6 Heretic artifacts2 measured conversions11.1K downloads
Best version97.8Pareto
trohrbaugh/Agents-A1-4B-hereticby trohrbaugh · Heretic v1.2.0 · MLX, SAFETENSORS
- KL divergence
- 0.045
- Refusals
- 0/100
- Refusal rate
- 0%
- Downloads
- 469
Evidence from the model card
| KL divergence | 0.0451 | | Refusals | 0/100 |
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 | trohrbaugh | 4B | 0.045 | 0/1000% | 97.8 | BestParetoMLXSAFETENSORS | 469 |
| 2 | richardyoung | 4B | 0.060 | 11/10011% | 91.7 | GGUFSAFETENSORS | 1.9K |
All artifacts (6)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| Agents-A1-4B-Heretic-ARA-Refusals8-GGUF NullpoLab | quant | GGUFBF16 | noneKL 0.0002 | 6.5K | 2026-07-22 |
| Agents-A1-4B-Heretic-ARA-Refusals0-GGUF NullpoLab | quant | GGUFBF16 | noneKL 0.012 | 1.8K | 2026-07-22 |
| Agents-A1-4B-heretic-GGUF richardyoung | quant | GGUFIQ3_M | published (weak match)KL 0.060 | 1.4K | 2026-10-15 |
| Agents-A1-4B-Heretic-ARA-Refusals7-GGUF NullpoLab | quant | GGUFBF16 | noneKL 0.0030 | 539 | 2026-07-22 |
| Agents-A1-4B-heretic richardyoung | heretic | SAFETENSORS | published (weak match)KL 0.060 | 496 | 2026-10-15 |
| Agents-A1-4B-heretic trohrbaugh | heretic | SAFETENSORS | published (weak match)KL 0.045 | 469 | 2026-08-06 |
3 artifacts inherit metrics from their parent conversion or publish none.