Llama 3 8B Slerp Med Chinese: benchmark results
shanchen's llama3-8B-slerp-med-chinese, a LazyMergekit SLERP merge of WiNGPT2-Llama-3-8B-Base and JSL-MedLlama-3-8B-v1.0, two Chinese/medical Llama 3 models. Provider: shanchen. Released 2024-04-30. Access: Open.
Aggregate Rating 1396 ± 28, rank #6298 of 10916 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Medical-LLM Leaderboard (BenchmarkList) | 72.34 | Medical average (%) | 90 |
| Open LLM Leaderboard v1 - MMLU | 64.96 | Accuracy (%) (5-shot) | 79.3 |
| Open Chinese LLM - CMMLU | 54.34 | Accuracy (%) | 56.5 |
| Open LLM Leaderboard v1 - GSM8K | 39.2 | Accuracy (%) (5-shot) | 54.1 |
| Open Chinese LLM - C-Eval Semantic | 68.24 | Accuracy (%) | 45.3 |
| Open Chinese LLM - HellaSwag | 59.12 | Accuracy (%) | 43.2 |
| Open LLM Leaderboard v1 - HellaSwag | 79.35 | Normalized accuracy (%) (10-shot) | 37.6 |
| Open LLM Leaderboard v1 - WinoGrande | 73.01 | Accuracy (%) (5-shot) | 30.7 |
| Open LLM Leaderboard v1 - ARC Challenge | 53.5 | Normalized accuracy (%) (25-shot) | 29.6 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 43.95 | MC2 (%) (0-shot) | 25.3 |
| Open Chinese LLM Leaderboard | 52.39 | Official Nonzero-Task Mean (%) | 24.6 |
| Open Chinese LLM - GSM8K | 30.86 | Accuracy (%) | 23.4 |
Interactive version: theaggregate.ai/model?slug=llama-3-8b-slerp-med-chinese · How It Works · Data refreshed daily, snapshot 2026-11-08.