Multimodal Commonsense Reasoning on M3CoT social-science and social-commonsense sub-topics
11.99Accuracy ChangeGPT-4o
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-4oPrompting Strategy=COCOT2025.07 | 11.99 | — | |
| Claude-3.5-SonnetPrompting Strategy=COCOT2025.07 | 5.9 | — | |
| GPT-4oPrompting Strategy=CoT2025.07 | 1.89 | — | |
| Gemini-3.0-FlashPrompting Strategy=COCOT, Model Type=Reasoning-First, Thinking mode=true2025.07 | 0.92 | — | |
| GPT-o1Prompting Strategy=COCOT, Model Type=Reasoning-First2025.07 | 0.06 | — | |
| Gemini-3.0-FlashPrompting Strategy=CoT, Model Type=Reasoning-First, Thinking mode=true2025.07 | -0.59 | — | |
| Claude-3.5-SonnetPrompting Strategy=CoT2025.07 | -2.28 | — | |
| GPT-4oPrompting Strategy=CCoT2025.07 | -7.23 | — | |
| Claude-3.5-SonnetPrompting Strategy=CCoT2025.07 | -9.78 | — | |
| Gemini-3.0-FlashPrompting Strategy=CCoT, Model Type=Reasoning-First, Thinking mode=true2025.07 | -11.3 | — | |
| GPT-o1Prompting Strategy=CoT, Model Type=Reasoning-First2025.07 | -15.15 | — | |
| GPT-o1Prompting Strategy=CCoT, Model Type=Reasoning-First2025.07 | -16.08 | — | |
| Claude-3.5-SonnetPrompting Strategy=Direct2025.07 | — | 65.1 | |
| Gemini-3.0-FlashPrompting Strategy=Direct, Model Type=Reasoning-First, Thinking mode=true2025.07 | — | 81.76 | |
| GPT-4oPrompting Strategy=Direct2025.07 | — | 57.23 | |
| GPT-o1Prompting Strategy=Direct, Model Type=Reasoning-First2025.07 | — | 81.4 |