List your company

Medical Visual Question Answering on Pooled medical VQA dataset Open

6.8ACELLaVA-NeXT-7B
4.91617.63330.3543.067Apr 2, 2026

Evaluation results

MethodLinks
LLaVA-NeXT-7BCalibration=Isotonic Regr., Confidence=Sampling2026.046.8
Qwen3-VL-8BCalibration=HAC-Platt, Confidence=Verbalized2026.047.5
Qwen3-VL-8BCalibration=HAC-Gate, Confidence=Verbalized2026.047.5
LLaVA-NeXT-7BCalibration=HAC-Gate, Confidence=Sampling2026.047.7
LLaVA-NeXT-7BCalibration=Platt Scaling, Confidence=Sampling2026.048.2
LLaVA-NeXT-7BCalibration=HAC-Platt, Confidence=Verbalized2026.048.3
LLaVA-NeXT-7BCalibration=HAC-Gate, Confidence=Verbalized2026.048.4
Qwen3-VL-8BCalibration=HAC-Platt, Confidence=Sampling2026.048.6
Qwen3-VL-8BCalibration=HAC-Gate, Confidence=Sampling2026.048.6
Qwen3-VL-8BCalibration=Isotonic Regr., Confidence=Verbalized2026.048.7
InternVL3-8BCalibration=HAC-Platt, Confidence=Verbalized2026.048.7
InternVL3-8BCalibration=Isotonic Regr., Confidence=Sampling2026.049
InternVL3-8BCalibration=HAC-Gate, Confidence=Verbalized2026.049.1
Qwen3-VL-8BCalibration=Isotonic Regr., Confidence=Sampling2026.049.4
InternVL3-8BCalibration=Platt Scaling, Confidence=Sampling2026.049.4
InternVL3-8BCalibration=Isotonic Regr., Confidence=Verbalized2026.049.4
InternVL3-8BCalibration=HAC-Platt, Confidence=Sampling2026.049.4
Qwen3-VL-8BCalibration=Platt Scaling, Confidence=Verbalized2026.049.5
LLaVA-NeXT-7BCalibration=HAC-Platt, Confidence=Sampling2026.049.6
InternVL3-8BCalibration=HAC-Gate, Confidence=Sampling2026.049.7
Qwen3-VL-8BCalibration=Platt Scaling, Confidence=Sampling2026.049.8
InternVL3-8BCalibration=Platt Scaling, Confidence=Verbalized2026.0410.2
LLaVA-NeXT-7BCalibration=Platt Scaling, Confidence=Verbalized2026.0410.4
LLaVA-NeXT-7BCalibration=Isotonic Regr., Confidence=Verbalized2026.0410.6
InternVL3-8BCalibration=Uncalibrated, Confidence=Sampling2026.0422.8
LLaVA-NeXT-7BCalibration=Uncalibrated, Confidence=Sampling2026.0425.6
InternVL3-8BCalibration=Uncalibrated, Confidence=Verbalized2026.0435.2
Qwen3-VL-8BCalibration=Uncalibrated, Confidence=Verbalized2026.0436.3
Qwen3-VL-8BCalibration=Uncalibrated, Confidence=Sampling2026.0440
LLaVA-NeXT-7BCalibration=Uncalibrated, Confidence=Verbalized2026.0453.9