LLM Evaluation on Chinese FuseEval
56.77Win RateSpecEM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SpecEMModel type=Methods of Ensembling Base LLMs, Reference model=Qwen2-72b-instruct2024.12 | 56.77 | 54.52 | |
| MOAModel type=Methods of Ensembling Base LLMs, Reference model=Qwen2-72b-instruct2024.12 | 53.12 | 51.24 | |
| GenFuseModel type=Methods of Ensembling Base LLMs, Reference model=Qwen2-72b-instruct2024.12 | 52.41 | 50.84 | |
| Qwen2.5-32b-instructModel type=Base LLMs, Reference model=Qwen2-72b-instruct2024.12 | 37.48 | 43.54 | |
| Mistral-24b-instruct-2501Model type=Base LLMs, Reference model=Qwen2-72b-instruct2024.12 | 31.79 | 44.27 | |
| UniTEModel type=Methods of Ensembling Base LLMs, Reference model=Qwen2-72b-instruct2024.12 | 19.13 | 41.04 | |
| Llama3-70b-instructModel type=Base LLMs, Reference model=Qwen2-72b-instruct2024.12 | 10.66 | 29.39 |