DeepSeek-R1-0528-Qwen3-8B
deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
8BQwen 3.87 Heretic artifacts3 measured conversions10.1K downloads
Best version98.1Pareto
saidutta69/DeepSeek-R1-Qwen3-8B-hereticby saidutta69 · Heretic v1.4.0 · GGUF, SAFETENSORS
- KL divergence
- 0.0082
- Refusals
- 3/100
- Refusal rate
- 3%
- Downloads
- 1.5K
Evidence from the model card
| KL divergence | 0.0082 | | Refusals | 3/100 |
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 | saidutta69 | 8B | 0.0082 | 3/1003% | 98.1 | BestParetoGGUFSAFETENSORS | 1.5K |
| 2 | richardyoung | 8B | 0.041 | 4/1004% | 96.0 | GGUFSAFETENSORS | 7.7K |
| 3 | jpvan4 | 8B | 0 | 60/10060% | 70.0 | unverifiedimplausibleParetoGGUFSAFETENSORS | 881 |
All artifacts (7)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| DeepSeek-R1-0528-Qwen3-8B-heretic-i1-GGUF mradermacher | quant | GGUFIQ1_M | inheritedKL 0.041 | 3.9K | 2026-09-30 |
| DeepSeek-R1-0528-Qwen3-8B-heretic-GGUF richardyoung | quant | GGUFIQ3_M | published (weak match)KL 0.041 | 2.1K | 2026-10-15 |
| DeepSeek-R1-Qwen3-8B-heretic saidutta69 | heretic | GGUFSAFETENSORSF16 | published (weak match)KL 0.0082 | 1.5K | 2026-10-08 |
| DeepSeek-R1-0528-Qwen3-8B-heretic-GGUF mradermacher | quant | GGUFIQ4_XS | inheritedKL 0.041 | 1K | 2026-09-30 |
| DeepSeek-R1-0528-Qwen3-8B-heretic richardyoung | heretic | SAFETENSORS | published (weak match)KL 0.041 | 682 | 2026-10-15 |
| DeepSeek-R1-0528-Qwen3-8B-heretic jpvan4 | heretic | SAFETENSORS | published (weak match)KL 0 | 669 | 2026-01-05 |
| DeepSeek-R1-0528-Qwen3-8B-heretic-Q4_K_M-GGUF wy8edhuh | quant | GGUFQ4_K_M | inheritedKL 0 | 212 | 2026-09-15 |