General Question Answering on PopQA (test val)
52.1Exact Match (EM)DeepControl
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DeepControlBackbone Model=Qwen2.5-7B2026.02 | 52.1 | 47.9 | |
| DeepControlBackbone Model=Qwen2.5-3B2026.02 | 51.2 | 41.1 | |
| Direct InferenceBackbone Model=Qwen2.5-7B2026.02 | 14 | 18.1 | |
| Direct InferenceBackbone Model=Qwen2.5-3B2026.02 | 10.8 | 13.4 |