ColorShadow-7B-v2: benchmark results
A Mistral-family gradient SLERP merge by nlpguy. The release recipe specifies layer-dependent attention and feed-forward interpolation; the numbered versions use distinct recipes. Provider: nlpguy. Access: Open.
Aggregate Rating 1412 ± 36, rank #5528 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 62.93 | MC2 (%) (0-shot) | 79.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 67.15 | Normalized accuracy (%) (25-shot) | 75.9 |
| Open LLM Leaderboard v1 - HellaSwag | 84.69 | Normalized accuracy (%) (10-shot) | 67.8 |
| Open LLM Leaderboard v1 - WinoGrande | 78.85 | Accuracy (%) (5-shot) | 65.9 |
| Open LLM Leaderboard v1 - GSM8K | 47.31 | Accuracy (%) (5-shot) | 60.7 |
| Open LLM Leaderboard v1 - MMLU | 60.34 | Accuracy (%) (5-shot) | 47.8 |
Interactive version: theaggregate.ai/model?slug=colorshadow-7b-v2 · How It Works · Data refreshed daily, snapshot 2026-11-08.