Bielik-11B-v3.0-Instruct
speakleash/Bielik-11B-v3.0-Instruct
11BBielik2 Heretic artifacts1 measured conversions2.5K downloads
Best version95.8Pareto
gepardzik/Bielik-11B-v3.0-Instruct-heretic-MPOAby gepardzik · Heretic v1.2.0 · GGUF, SAFETENSORS
- KL divergence
- 0.036
- Refusals
- 5/100
- Refusal rate
- 5%
- Downloads
- 2.5K
Evidence from the model card
| KL divergence | 0.0356 | | Refusals | 5/100 |
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 | gepardzik | 11B | 0.036 | 5/1005% | 95.8 | BestParetoGGUFSAFETENSORS | 2.5K |
All artifacts (2)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| Bielik-11B-v3.0-Instruct-heretic-MPOA gepardzik | heretic | SAFETENSORS | published (weak match)KL 0.036 | 1.3K | 2026-02-23 |
| Bielik-11B-v3.0-Instruct-heretic-MPOA-GGUF gepardzik | quant | GGUF | inheritedKL 0.036 | 1.2K | 2026-02-23 |