Visual Math & Hallucination on HallusionBench
76AccuracyPDCR
Evaluation Results
| Method | Links | |
|---|---|---|
| PDCRBackbone=Qwen3-VL-8B-Instruct2026.05 | 76 | |
| DAPOBackbone=Qwen3-VL-8B-Instruct2026.05 | 75.8 | |
| GRPOBackbone=Qwen3-VL-8B-Instruct2026.05 | 75.6 | |
| PACRBackbone=Qwen3-VL-8B-Instruct2026.05 | 75.5 | |
| Zero-shot InferenceBackbone=Qwen3-VL-8B-Instruct2026.05 | 74.9 |