Medical Visual Question Answering on SLAKE Closed
36.7ACELLaVA-NeXT-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaVA-NeXT-7BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 36.7 | |
| LLaVA-NeXT-34BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 24.5 | |
| LLaVA-NeXT-7BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 23.5 | |
| LLaVA-NeXT-34BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 21.7 | |
| LLaVA-NeXT-7BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 20.8 | |
| Qwen3-VL-8BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 19.8 | |
| Qwen3-VL-32BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 19.7 | |
| InternVL3-8BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 19.7 | |
| LLaVA-NeXT-7BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 19.7 | |
| LLaVA-NeXT-7BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 19.6 | |
| InternVL3-2BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 19.4 | |
| Qwen3-VL-2BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 18.9 | |
| LLaVA-NeXT-7BCalibration=Hist. Binning, Confidence=Sampling2026.04 | 18.8 | |
| LLaVA-NeXT-7BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 18.6 | |
| InternVL3-2BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 18.5 | |
| InternVL3-2BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 18.1 | |
| LLaVA-NeXT-7BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 18.1 | |
| LLaVA-NeXT-7BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 18.1 | |
| LLaVA-NeXT-7BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 18.1 | |
| Qwen3-VL-8BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 18 | |
| InternVL3-2BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 17.6 | |
| InternVL3-38BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 17.5 | |
| LLaVA-NeXT-34BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 17.3 | |
| Qwen3-VL-2BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 17.2 | |
| InternVL3-2BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 17.2 | |
| LLaVA-NeXT-7BCalibration=Hist. Binning, Confidence=Verbalized2026.04 | 17.1 | |
| LLaVA-NeXT-34BCalibration=Hist. Binning, Confidence=Verbalized2026.04 | 17.1 | |
| Qwen3-VL-32BCalibration=Hist. Binning, Confidence=Verbalized2026.04 | 17 | |
| LLaVA-NeXT-34BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 17 | |
| LLaVA-NeXT-34BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 16.9 | |
| InternVL3-2BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 16.7 | |
| Qwen3-VL-32BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 16.6 | |
| Qwen3-VL-32BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 16.6 | |
| InternVL3-8BCalibration=Hist. Binning, Confidence=Sampling2026.04 | 16.4 | |
| Qwen3-VL-32BCalibration=Uncalibrated, Confidence=Verbalized2026.04 | 16.3 | |
| LLaVA-NeXT-7BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 16.2 | |
| InternVL3-2BCalibration=Hist. Binning, Confidence=Sampling2026.04 | 16.1 | |
| InternVL3-38BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 15.8 | |
| LLaVA-NeXT-34BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 15.8 | |
| InternVL3-2BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 15.7 | |
| InternVL3-8BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 15.7 | |
| InternVL3-2BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 15.6 | |
| InternVL3-8BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 15.6 | |
| InternVL3-2BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 15.5 | |
| LLaVA-NeXT-34BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 15.4 | |
| InternVL3-8BCalibration=Hist. Binning, Confidence=Verbalized2026.04 | 15.3 | |
| InternVL3-38BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 15.3 | |
| LLaVA-NeXT-34BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 15.3 | |
| InternVL3-8BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 15.2 | |
| InternVL3-8BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 15.2 | |
| InternVL3-8BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 15.2 | |
| InternVL3-8BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 15.2 | |
| LLaVA-NeXT-34BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 15.2 | |
| InternVL3-2BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 15.1 | |
| InternVL3-2BCalibration=Hist. Binning, Confidence=Verbalized2026.04 | 15.1 | |
| InternVL3-38BCalibration=Hist. Binning, Confidence=Verbalized2026.04 | 15.1 | |
| Qwen3-VL-8BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 14.8 | |
| InternVL3-8BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 14.8 | |
| Qwen3-VL-2BCalibration=Hist. Binning, Confidence=Sampling2026.04 | 14.6 | |
| InternVL3-38BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 14.6 | |
| Qwen3-VL-8BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 14.4 | |
| InternVL3-8BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 14.1 | |
| LLaVA-NeXT-34BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 14.1 | |
| Qwen3-VL-2BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 13.9 | |
| Qwen3-VL-2BCalibration=HAC-Gate, Confidence=Verbalized2026.04 | 13.8 | |
| Qwen3-VL-8BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 13.8 | |
| Qwen3-VL-2BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 13.7 | |
| Qwen3-VL-2BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 13.7 | |
| Qwen3-VL-32BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 13.7 | |
| Qwen3-VL-32BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 13.7 | |
| Qwen3-VL-32BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 13.7 | |
| InternVL3-8BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 13.6 | |
| LLaVA-NeXT-34BCalibration=Hist. Binning, Confidence=Sampling2026.04 | 13.6 | |
| Qwen3-VL-2BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 13.3 | |
| Qwen3-VL-8BCalibration=Hist. Binning, Confidence=Verbalized2026.04 | 13.3 | |
| Qwen3-VL-8BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 13.2 | |
| Qwen3-VL-8BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 13.2 | |
| Qwen3-VL-2BCalibration=HAC-Platt, Confidence=Verbalized2026.04 | 13.1 | |
| Qwen3-VL-8BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 13 | |
| Qwen3-VL-8BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 12.9 | |
| InternVL3-38BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 12.8 | |
| Qwen3-VL-8BCalibration=Hist. Binning, Confidence=Sampling2026.04 | 12.7 | |
| InternVL3-38BCalibration=Hist. Binning, Confidence=Sampling2026.04 | 12.7 | |
| Qwen3-VL-2BCalibration=Platt Scaling, Confidence=Verbalized2026.04 | 12.5 | |
| Qwen3-VL-2BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 12.4 | |
| InternVL3-38BCalibration=Uncalibrated, Confidence=Sampling2026.04 | 12 | |
| InternVL3-38BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 12 | |
| InternVL3-38BCalibration=HAC-Gate, Confidence=Sampling2026.04 | 11.9 | |
| Qwen3-VL-2BCalibration=Hist. Binning, Confidence=Verbalized2026.04 | 11.7 | |
| Qwen3-VL-8BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 11.7 | |
| Qwen3-VL-32BCalibration=Isotonic Regr., Confidence=Verbalized2026.04 | 11.4 | |
| InternVL3-38BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 11.2 | |
| InternVL3-38BCalibration=HAC-Platt, Confidence=Sampling2026.04 | 11.2 | |
| Qwen3-VL-32BCalibration=Platt Scaling, Confidence=Sampling2026.04 | 11.1 | |
| Qwen3-VL-32BCalibration=Isotonic Regr., Confidence=Sampling2026.04 | 11.1 | |
| Qwen3-VL-32BCalibration=Hist. Binning, Confidence=Sampling2026.04 | 11.1 |