Noromaid-7B-0.4-DPO [float16] [rev: 47a417a21670]: benchmark results
A Noromaid full fine-tune with DPO by NeverSleep, credited to Undi and IkariDev. The card specifies ChatML prompting and DPO datasets including Intel/orca_dpo_pairs, ToxicDPOqa and toxic-dpo-v0.1-NoWarning. Provider: NeverSleep. Access: Open.
Aggregate Rating 1398 ± 34, rank #6212 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - HellaSwag | 84.32 | Normalized accuracy (%) (10-shot) | 65.4 |
| Open LLM Leaderboard v1 - MMLU | 63.2 | Accuracy (%) (5-shot) | 58.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 62.29 | Normalized accuracy (%) (25-shot) | 57.1 |
| Open LLM Leaderboard v1 - WinoGrande | 76.95 | Accuracy (%) (5-shot) | 52 |
| Open LLM Leaderboard v1 - GSM8K | 25.47 | Accuracy (%) (5-shot) | 44 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 42.28 | MC2 (%) (0-shot) | 20.1 |
Interactive version: theaggregate.ai/model?slug=noromaid-7b-0-4-dpo-float16-rev-47a417a21670 · How It Works · Data refreshed daily, snapshot 2026-11-08.