NeuralDaredevil-8B-abliterated: benchmark results
mlabonne's abliterated Daredevil-8B Llama 3 merge, DPO-tuned on orpo-dpo-mix-40k to recover the performance lost to abliteration. Provider: mlabonne. Released 2024-05-27. Access: Open.
Aggregate Rating 1469 ± 20, rank #892 of 1656 rated models, from 52 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - IFEval | 75.61 | Score | 91.2 |
| Open PL LLM - PSC (generative, 5-shot) | 97.94 | Binary F1 (%) | 90.9 |
| Open PL LLM - PolEmo2-IN (multiple choice, 5-shot) | 84.35 | Accuracy (%) | 83.1 |
| Open Portuguese LLM - ASSIN2 RTE | 92.73 | Macro F1 (%) | 82.8 |
| Open PL LLM - KLEJ NER (multiple choice, 5-shot) | 54.37 | Accuracy (%) | 79.5 |
| Open Portuguese LLM - ENEM | 71.17 | Accuracy (%) | 79.3 |
| Open Portuguese LLM - BLUEX | 60.22 | Accuracy (%) | 78 |
| Open Portuguese LLM - OAB Exams | 50.48 | Accuracy (%) | 77.2 |
| Open PL LLM - CBD (multiple choice, 5-shot) | 32.83 | Macro F1 (%) | 75.5 |
| Open Chinese LLM - GSM8K | 53.83 | Accuracy (%) | 73.2 |
| Open PL LLM - PolEmo2-IN (generative, 5-shot) | 81.3 | Accuracy (%) | 72 |
| Open PL LLM - PoQuAD Open Book (generative, 5-shot) | 67.49 | Levenshtein Similarity (%) | 71.8 |
Interactive version: theaggregate.ai/model?slug=neuraldaredevil-8b-abliterated · How It Works · Data refreshed daily, snapshot 2026-10-20.