Question Answering on RealEHR
55.2AccuracyGATE
Evaluation Results
| Method | Links | |
|---|---|---|
| GATEBackbone LLM=GPT-5.4-mini2026.06 | 55.2 | |
| ReFoRCEBackbone LLM=GPT-5.4-mini2026.06 | 48.6 | |
| ReDelBackbone LLM=GPT-5.4-mini2026.06 | 44.8 | |
| ReActBackbone LLM=GPT-5.4-mini2026.06 | 42.9 | |
| GATEBackbone LLM=Qwen3-8B2026.06 | 5.6 | |
| ReFoRCEBackbone LLM=Qwen3-8B2026.06 | 3.8 | |
| ReDelBackbone LLM=Qwen3-8B2026.06 | 2.5 | |
| ReActBackbone LLM=Qwen3-8B2026.06 | 1.7 |