Medical Visual Question Answering on SLAKE (0-shot vs CoT Deltas)
63.8AccuracyOctoTools
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| OctoToolsBackbone=gpt-4o-2024-08-06, Modality=Vision, Domain=Medical, Visual Understanding=true, Numerical Calculation=false, Knowledge Retrieval=true, Multi-step Reasoning=true2025.02 | 63.8 | 4.5 | 3.5 | |
| CoTBackbone=gpt-4o-2024-08-06, Modality=Vision, Domain=Medical, Visual Understanding=true, Numerical Calculation=false, Knowledge Retrieval=true, Multi-step Reasoning=true2025.02 | 60.3 | — | — | |
| 0-shotBackbone=gpt-4o-2024-08-06, Modality=Vision, Domain=Medical, Visual Understanding=true, Numerical Calculation=false, Knowledge Retrieval=true, Multi-step Reasoning=true2025.02 | 59.3 | — | — | |
| OctoToolsbaseBackbone=gpt-4o-2024-08-06, Modality=Vision, Domain=Medical, Visual Understanding=true, Numerical Calculation=false, Knowledge Retrieval=true, Multi-step Reasoning=true2025.02 | 59.2 | — | — |