Visual Question Answering on AlgoPuzzleVQA
48.7AccuracyOctoTools
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| OctoToolsBackbone=gpt-4o-2024-08-06, Modality=Vision, Domain=General, Visual Understanding=true, Numerical Calculation=false, Knowledge Retrieval=false, Multi-step Reasoning=true2025.02 | 48.7 | 7.4 | 6 | |
| OctoToolsbaseBackbone=gpt-4o-2024-08-06, Modality=Vision, Domain=General, Visual Understanding=true, Numerical Calculation=false, Knowledge Retrieval=false, Multi-step Reasoning=true2025.02 | 44 | — | — | |
| CoTBackbone=gpt-4o-2024-08-06, Modality=Vision, Domain=General, Visual Understanding=true, Numerical Calculation=false, Knowledge Retrieval=false, Multi-step Reasoning=true2025.02 | 42.7 | — | — | |
| 0-shotBackbone=gpt-4o-2024-08-06, Modality=Vision, Domain=General, Visual Understanding=true, Numerical Calculation=false, Knowledge Retrieval=false, Multi-step Reasoning=true2025.02 | 41.3 | — | — |