Scientific Question Answering on GPQA (Accuracy, Delta 0-shot, Delta CoT)
54.7AccuracyOctoTools
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| OctoToolsBackbone=gpt-4o-2024-08-06, Modality=Text, Domain=Scientific, Visual Understanding=false, Numerical Calculation=true, Knowledge Retrieval=true, Multi-step Reasoning=true2025.02 | 54.7 | 1 | 2.4 | |
| 0-shotBackbone=gpt-4o-2024-08-06, Modality=Text, Domain=Scientific, Visual Understanding=false, Numerical Calculation=true, Knowledge Retrieval=true, Multi-step Reasoning=true2025.02 | 53.7 | — | — | |
| OctoToolsbaseBackbone=gpt-4o-2024-08-06, Modality=Text, Domain=Scientific, Visual Understanding=false, Numerical Calculation=true, Knowledge Retrieval=true, Multi-step Reasoning=true2025.02 | 53.7 | — | — | |
| CoTBackbone=gpt-4o-2024-08-06, Modality=Text, Domain=Scientific, Visual Understanding=false, Numerical Calculation=true, Knowledge Retrieval=true, Multi-step Reasoning=true2025.02 | 52.3 | — | — |