Llama-3.2-1B-Instruct-Capybara
joelewing/Llama-3.2-1B-Instruct-Capybara
1BLlama 3.22 Heretic artifacts1 measured conversions309 downloads
Best version97.0Pareto
joelewing/Llama-3.2-1B-Instruct-Capybara-abliterated-GGUFby joelewing · Heretic v1.0.1 · GGUF, SAFETENSORS
- KL divergence
- 0.030
- Refusals
- 3/100
- Refusal rate
- 3%
- Downloads
- 309
Evidence from the model card
| KL divergence | 0.03 | | 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 | joelewing | 1B | 0.030 | 3/1003% | 97.0 | BestParetoGGUFSAFETENSORS | 309 |
All artifacts (2)
| Repository | Type | Formats | Metrics | Downloads | Updated |
|---|---|---|---|---|---|
| Llama-3.2-1B-Instruct-Capybara-abliterated-GGUF joelewing | heretic | GGUFQ8_0 | published (weak match)KL 0.030 | 168 | 2025-11-17 |
| Llama-3.2-1B-Instruct-Capybara-abliterated joelewing | heretic | SAFETENSORS | published (weak match)KL 0.030 | 141 | 2025-11-17 |