Meta-Llama-3-8B-Instruct
meta-llama/Meta-Llama-3-8B-Instruct
8BLlama 31 Heretic artifacts1 measured conversions13.9K downloads
Best version93.3Pareto
lenalibon/Meta-Llama-3-8B-Instruct-heretic-mlabonneby lenalibon · Heretic v1.2.0 · SAFETENSORS
- KL divergence
- 0.080
- Refusals
- 6/100
- Refusal rate
- 6%
- Downloads
- 13.9K
Evidence from the model card
| KL divergence | 0.0799 | | Refusals | 6/100 |
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 | lenalibon | 8B | 0.080 | 6/1006% | 93.3 | BestParetoSAFETENSORS | 13.9K |
All artifacts (1)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| Meta-Llama-3-8B-Instruct-heretic-mlabonne lenalibon | heretic | SAFETENSORS | published (weak match)KL 0.080 | 13.9K | 2026-03-10 |