List your company

Medical Visual Question Answering on SLAKE Closed

36.7ACELLaVA-NeXT-7B
10.07616.98823.930.812Apr 2, 2026

Evaluation results

MethodLinks
LLaVA-NeXT-7BCalibration=Uncalibrated, Confidence=Verbalized2026.0436.7
LLaVA-NeXT-34BCalibration=Uncalibrated, Confidence=Verbalized2026.0424.5
LLaVA-NeXT-7BCalibration=HAC-Gate, Confidence=Verbalized2026.0423.5
LLaVA-NeXT-34BCalibration=Uncalibrated, Confidence=Sampling2026.0421.7
LLaVA-NeXT-7BCalibration=Isotonic Regr., Confidence=Verbalized2026.0420.8
Qwen3-VL-8BCalibration=Uncalibrated, Confidence=Sampling2026.0419.8
Qwen3-VL-32BCalibration=Uncalibrated, Confidence=Sampling2026.0419.7
InternVL3-8BCalibration=Uncalibrated, Confidence=Verbalized2026.0419.7
LLaVA-NeXT-7BCalibration=Platt Scaling, Confidence=Verbalized2026.0419.7
LLaVA-NeXT-7BCalibration=HAC-Platt, Confidence=Verbalized2026.0419.6
InternVL3-2BCalibration=Uncalibrated, Confidence=Verbalized2026.0419.4
Qwen3-VL-2BCalibration=Uncalibrated, Confidence=Sampling2026.0418.9
LLaVA-NeXT-7BCalibration=Hist. Binning, Confidence=Sampling2026.0418.8
LLaVA-NeXT-7BCalibration=Uncalibrated, Confidence=Sampling2026.0418.6
InternVL3-2BCalibration=HAC-Platt, Confidence=Sampling2026.0418.5
InternVL3-2BCalibration=Uncalibrated, Confidence=Sampling2026.0418.1
LLaVA-NeXT-7BCalibration=Isotonic Regr., Confidence=Sampling2026.0418.1
LLaVA-NeXT-7BCalibration=HAC-Platt, Confidence=Sampling2026.0418.1
LLaVA-NeXT-7BCalibration=HAC-Gate, Confidence=Sampling2026.0418.1
Qwen3-VL-8BCalibration=Uncalibrated, Confidence=Verbalized2026.0418
InternVL3-2BCalibration=Platt Scaling, Confidence=Sampling2026.0417.6
InternVL3-38BCalibration=Uncalibrated, Confidence=Verbalized2026.0417.5
LLaVA-NeXT-34BCalibration=Isotonic Regr., Confidence=Verbalized2026.0417.3
Qwen3-VL-2BCalibration=Uncalibrated, Confidence=Verbalized2026.0417.2
InternVL3-2BCalibration=Isotonic Regr., Confidence=Verbalized2026.0417.2
LLaVA-NeXT-7BCalibration=Hist. Binning, Confidence=Verbalized2026.0417.1
LLaVA-NeXT-34BCalibration=Hist. Binning, Confidence=Verbalized2026.0417.1
Qwen3-VL-32BCalibration=Hist. Binning, Confidence=Verbalized2026.0417
LLaVA-NeXT-34BCalibration=HAC-Platt, Confidence=Sampling2026.0417
LLaVA-NeXT-34BCalibration=Platt Scaling, Confidence=Verbalized2026.0416.9
InternVL3-2BCalibration=HAC-Gate, Confidence=Sampling2026.0416.7
Qwen3-VL-32BCalibration=HAC-Platt, Confidence=Verbalized2026.0416.6
Qwen3-VL-32BCalibration=HAC-Gate, Confidence=Verbalized2026.0416.6
InternVL3-8BCalibration=Hist. Binning, Confidence=Sampling2026.0416.4
Qwen3-VL-32BCalibration=Uncalibrated, Confidence=Verbalized2026.0416.3
LLaVA-NeXT-7BCalibration=Platt Scaling, Confidence=Sampling2026.0416.2
InternVL3-2BCalibration=Hist. Binning, Confidence=Sampling2026.0416.1
InternVL3-38BCalibration=HAC-Gate, Confidence=Verbalized2026.0415.8
LLaVA-NeXT-34BCalibration=HAC-Gate, Confidence=Verbalized2026.0415.8
InternVL3-2BCalibration=HAC-Gate, Confidence=Verbalized2026.0415.7
InternVL3-8BCalibration=Isotonic Regr., Confidence=Verbalized2026.0415.7
InternVL3-2BCalibration=HAC-Platt, Confidence=Verbalized2026.0415.6
InternVL3-8BCalibration=Platt Scaling, Confidence=Sampling2026.0415.6
InternVL3-2BCalibration=Isotonic Regr., Confidence=Sampling2026.0415.5
LLaVA-NeXT-34BCalibration=HAC-Platt, Confidence=Verbalized2026.0415.4
InternVL3-8BCalibration=Hist. Binning, Confidence=Verbalized2026.0415.3
InternVL3-38BCalibration=HAC-Platt, Confidence=Verbalized2026.0415.3
LLaVA-NeXT-34BCalibration=Platt Scaling, Confidence=Sampling2026.0415.3
InternVL3-8BCalibration=Uncalibrated, Confidence=Sampling2026.0415.2
InternVL3-8BCalibration=Platt Scaling, Confidence=Verbalized2026.0415.2
InternVL3-8BCalibration=Isotonic Regr., Confidence=Sampling2026.0415.2
InternVL3-8BCalibration=HAC-Platt, Confidence=Sampling2026.0415.2
LLaVA-NeXT-34BCalibration=HAC-Gate, Confidence=Sampling2026.0415.2
InternVL3-2BCalibration=Platt Scaling, Confidence=Verbalized2026.0415.1
InternVL3-2BCalibration=Hist. Binning, Confidence=Verbalized2026.0415.1
InternVL3-38BCalibration=Hist. Binning, Confidence=Verbalized2026.0415.1
Qwen3-VL-8BCalibration=HAC-Platt, Confidence=Verbalized2026.0414.8
InternVL3-8BCalibration=HAC-Gate, Confidence=Sampling2026.0414.8
Qwen3-VL-2BCalibration=Hist. Binning, Confidence=Sampling2026.0414.6
InternVL3-38BCalibration=Platt Scaling, Confidence=Verbalized2026.0414.6
Qwen3-VL-8BCalibration=HAC-Gate, Confidence=Verbalized2026.0414.4
InternVL3-8BCalibration=HAC-Gate, Confidence=Verbalized2026.0414.1
LLaVA-NeXT-34BCalibration=Isotonic Regr., Confidence=Sampling2026.0414.1
Qwen3-VL-2BCalibration=HAC-Gate, Confidence=Sampling2026.0413.9
Qwen3-VL-2BCalibration=HAC-Gate, Confidence=Verbalized2026.0413.8
Qwen3-VL-8BCalibration=Platt Scaling, Confidence=Verbalized2026.0413.8
Qwen3-VL-2BCalibration=Isotonic Regr., Confidence=Sampling2026.0413.7
Qwen3-VL-2BCalibration=HAC-Platt, Confidence=Sampling2026.0413.7
Qwen3-VL-32BCalibration=Platt Scaling, Confidence=Verbalized2026.0413.7
Qwen3-VL-32BCalibration=HAC-Platt, Confidence=Sampling2026.0413.7
Qwen3-VL-32BCalibration=HAC-Gate, Confidence=Sampling2026.0413.7
InternVL3-8BCalibration=HAC-Platt, Confidence=Verbalized2026.0413.6
LLaVA-NeXT-34BCalibration=Hist. Binning, Confidence=Sampling2026.0413.6
Qwen3-VL-2BCalibration=Platt Scaling, Confidence=Sampling2026.0413.3
Qwen3-VL-8BCalibration=Hist. Binning, Confidence=Verbalized2026.0413.3
Qwen3-VL-8BCalibration=HAC-Platt, Confidence=Sampling2026.0413.2
Qwen3-VL-8BCalibration=HAC-Gate, Confidence=Sampling2026.0413.2
Qwen3-VL-2BCalibration=HAC-Platt, Confidence=Verbalized2026.0413.1
Qwen3-VL-8BCalibration=Platt Scaling, Confidence=Sampling2026.0413
Qwen3-VL-8BCalibration=Isotonic Regr., Confidence=Sampling2026.0412.9
InternVL3-38BCalibration=Platt Scaling, Confidence=Sampling2026.0412.8
Qwen3-VL-8BCalibration=Hist. Binning, Confidence=Sampling2026.0412.7
InternVL3-38BCalibration=Hist. Binning, Confidence=Sampling2026.0412.7
Qwen3-VL-2BCalibration=Platt Scaling, Confidence=Verbalized2026.0412.5
Qwen3-VL-2BCalibration=Isotonic Regr., Confidence=Verbalized2026.0412.4
InternVL3-38BCalibration=Uncalibrated, Confidence=Sampling2026.0412
InternVL3-38BCalibration=Isotonic Regr., Confidence=Verbalized2026.0412
InternVL3-38BCalibration=HAC-Gate, Confidence=Sampling2026.0411.9
Qwen3-VL-2BCalibration=Hist. Binning, Confidence=Verbalized2026.0411.7
Qwen3-VL-8BCalibration=Isotonic Regr., Confidence=Verbalized2026.0411.7
Qwen3-VL-32BCalibration=Isotonic Regr., Confidence=Verbalized2026.0411.4
InternVL3-38BCalibration=Isotonic Regr., Confidence=Sampling2026.0411.2
InternVL3-38BCalibration=HAC-Platt, Confidence=Sampling2026.0411.2
Qwen3-VL-32BCalibration=Platt Scaling, Confidence=Sampling2026.0411.1
Qwen3-VL-32BCalibration=Isotonic Regr., Confidence=Sampling2026.0411.1
Qwen3-VL-32BCalibration=Hist. Binning, Confidence=Sampling2026.0411.1