Qwen2.5-3B-instruct-medical-finetuned
Joker-sxj/Qwen2.5-3B-instruct-medical-finetuned
3BQwen 2.54 Heretic artifacts1 measured conversions8.3K downloads
Best version97.4Pareto
ChiKoi7/Qwen2.5-3B-instruct-medical-finetuned-Hereticby ChiKoi7 · Heretic v1.2.0 · GGUF, SAFETENSORS
- KL divergence
- 0.013
- Refusals
- 4/100
- Refusal rate
- 4%
- Downloads
- 7.3K
Evidence from the model card
| Refusals | 4/100 | | KL divergence | 0.0126 |
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 | ChiKoi7 | 3B | 0.013 | 4/1004% | 97.4 | BestParetoGGUFSAFETENSORS | 7.3K |
All artifacts (4)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| Qwen2.5-3B-instruct-medical-finetuned-Heretic-i1-GGUF mradermacher | quant | GGUFIQ1_M | inheritedKL 0.013 | 5.5K | 2026-04-22 |
| Qwen2.5-3B-instruct-medical-finetuned-Heretic-GGUF mradermacher | quant | GGUFIQ4_XS | inheritedKL 0.013 | 1.8K | 2026-04-22 |
| evolva-qwen25-3b-medical-heretic-GGUF mradermacher | quant | GGUFIQ4_XS | none | 988 | 2026-07-04 |
| Qwen2.5-3B-instruct-medical-finetuned-Heretic ChiKoi7 | heretic | SAFETENSORS | published (weak match)KL 0.013 | 71 | 2026-04-21 |
1 artifact inherit metrics from their parent conversion or publish none.