SHRDFU-7B-beta: benchmark results
maldv SHRDFU 7B Beta is a fine-tune of cgato/Thespis-CurtainCall-7b-v0.3 using the crabcanon dataset, with attention-layer PEFT conditioning followed by small full-layer tuning. Provider: maldv. Access: Open.
Aggregate Rating 1398 ± 35, rank #6224 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - ARC Challenge | 66.38 | Normalized accuracy (%) (25-shot) | 73.4 |
| Open LLM Leaderboard v1 - HellaSwag | 85.03 | Normalized accuracy (%) (10-shot) | 71.2 |
| Open LLM Leaderboard v1 - WinoGrande | 77.74 | Accuracy (%) (5-shot) | 58.2 |
| Open LLM Leaderboard v1 - MMLU | 60.29 | Accuracy (%) (5-shot) | 47.5 |
| Open LLM Leaderboard v1 - GSM8K | 27.9 | Accuracy (%) (5-shot) | 45.6 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 49.45 | MC2 (%) (0-shot) | 44.8 |
Interactive version: theaggregate.ai/model?slug=shrdfu-7b-beta · How It Works · Data refreshed daily, snapshot 2026-11-08.