Error Detection on CoSPlan Robo-VQA-E
45.3AccuracyGPT-4o
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-4oReasoning Strategy=Chain-of-Thought2025.12 | 45.3 | |
| GPT-4oReasoning Strategy=Scene Graph2025.12 | 44.2 | |
| CoG-VLMReasoning Strategy=Scene Graph2025.12 | 35.3 | |
| CoG-VLMReasoning Strategy=Chain-of-Thought2025.12 | 33.4 | |
| CoG-VLMReasoning Strategy=Vanilla2025.12 | 32.1 | |
| Janus-pro-7BReasoning Strategy=Scene Graph2025.12 | 26.1 | |
| Intern-VLMReasoning Strategy=Scene Graph2025.12 | 26.1 | |
| Random2025.12 | 25.4 | |
| Intern-VLMReasoning Strategy=Chain-of-Thought2025.12 | 25.2 | |
| Intern-VLMReasoning Strategy=Vanilla2025.12 | 24.3 | |
| Janus-pro-7BReasoning Strategy=Chain-of-Thought2025.12 | 18.1 | |
| Janus-pro-7BReasoning Strategy=Vanilla2025.12 | 17.5 | |
| Qwen2 VL-8BReasoning Strategy=Scene Graph2025.12 | 9.6 | |
| Qwen2 VL-8BReasoning Strategy=Vanilla2025.12 | 9.2 | |
| Qwen2 VL-8BReasoning Strategy=Chain-of-Thought2025.12 | 9.1 |