Wizard-Vicuna-13B-Uncensored-HF: benchmark results
TheBloke's Hugging Face publication of Eric Hartford's 'uncensored' training of Wizard-Vicuna 13B, whose original release is ehartford/Wizard-Vicuna-13B-Uncensored. Conversion publication is credited separately from the original model creator. Provider: TheBloke. Access: Open.
Aggregate Rating 1304 ± 15, rank #8658 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - TruthfulQA MC1 | 36.35 | Accuracy (%) | 87 |
| MMLU-by-task - TruthfulQA MC2 | 51.69 | Accuracy (%) | 84.3 |
| MMLU-by-task - ARC Challenge | 56.57 | Accuracy (%) | 77.3 |
| MMLU-by-task - HellaSwag | 62.19 | Accuracy (%) | 74.8 |
| MMLU-by-task - Machine Learning | 37.5 | Accuracy (%) | 71.9 |
| MMLU-by-task - Virology | 45.18 | Accuracy (%) | 70.3 |
| MMLU-by-task - Formal Logic | 34.92 | Accuracy (%) | 69.3 |
| MMLU-by-task - Medical Genetics | 57 | Accuracy (%) | 68.4 |
| MMLU-by-task - Professional Law | 40.55 | Accuracy (%) | 67.9 |
| MMLU-by-task - US Foreign Policy | 77 | Accuracy (%) | 65.8 |
| MMLU-by-task - Human Sexuality | 59.54 | Accuracy (%) | 63.9 |
| MMLU-by-task - High School Computer Science | 52 | Accuracy (%) | 63.6 |
Interactive version: theaggregate.ai/model?slug=wizard-vicuna-13b-uncensored-hf · How It Works · Data refreshed daily, snapshot 2026-11-08.