Medical Visual Question Answering on Pooled medical VQA dataset Open
6.8ACELLaVA-NeXT-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaVA-NeXT-7BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 6.8 | |
| Qwen3-VL-8BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 7.5 | |
| Qwen3-VL-8BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 7.5 | |
| LLaVA-NeXT-7BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 7.7 | |
| LLaVA-NeXT-7BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 8.2 | |
| LLaVA-NeXT-7BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 8.3 | |
| LLaVA-NeXT-7BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 8.4 | |
| Qwen3-VL-8BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 8.6 | |
| Qwen3-VL-8BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 8.6 | |
| Qwen3-VL-8BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 8.7 | |
| InternVL3-8BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 8.7 | |
| InternVL3-8BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 9 | |
| InternVL3-8BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 9.1 | |
| Qwen3-VL-8BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 9.4 | |
| InternVL3-8BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 9.4 | |
| InternVL3-8BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 9.4 | |
| InternVL3-8BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 9.4 | |
| Qwen3-VL-8BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 9.5 | |
| LLaVA-NeXT-7BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 9.6 | |
| InternVL3-8BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 9.7 | |
| Qwen3-VL-8BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 9.8 | |
| InternVL3-8BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 10.2 | |
| LLaVA-NeXT-7BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 10.4 | |
| LLaVA-NeXT-7BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 10.6 | |
| InternVL3-8BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 22.8 | |
| LLaVA-NeXT-7BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 25.6 | |
| InternVL3-8BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 35.2 | |
| Qwen3-VL-8BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 36.3 | |
| Qwen3-VL-8BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 40 | |
| LLaVA-NeXT-7BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 53.9 |