Command Understanding on MiCU Dataset
94.61Overall AccuracyMiCU-4B
Evaluation Results
| Method | Links | |
|---|---|---|
| MiCU-4BType=Open Source LLM2026.05 | 94.61 | |
| MiCU-4B-fastType=Open Source LLM2026.05 | 94.01 | |
| DeepSeek-V3.2Type=Proprietary LLM2026.05 | 74.6 | |
| GPT-4oType=Proprietary LLM2026.05 | 73.57 | |
| DeepSeek-R1Type=Proprietary LLM2026.05 | 73.1 | |
| Rule-based SelectionType=Rule-based2026.05 | 66.32 | |
| Qwen3-30BType=Open Source LLM2026.05 | 64.01 | |
| GPT-4o-miniType=Proprietary LLM2026.05 | 63.79 | |
| Llama3.3-70BType=Open Source LLM2026.05 | 61.61 | |
| Llama3.1-8BType=Open Source LLM2026.05 | 54.31 | |
| Qwen3-4BType=Open Source LLM2026.05 | 34.11 |