Task Planning on RS Task Planning Dataset
100Object Counting AccuracyRS-Agent
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| RS-AgentLLM Backbone=gpt-3.5-turbo2024.06 | 100 | 89.47 | 95.24 | 61.9 | 100 | 68.42 | 100 | 87.77 | |
| RS-ChatGPTLLM Backbone=gpt-4o-mini2024.06 | 100 | 42.11 | 90.48 | 100 | 100 | 89.47 | 94.74 | 88.49 | |
| RS-AgentLLM Backbone=gpt-4o-mini2024.06 | 100 | 94.74 | 90.48 | 95.24 | 90.48 | 100 | 100 | 95.68 | |
| RS-ChatGPTLLM Backbone=gpt-3.5-turbo2024.06 | 97.74 | 52.63 | 95.24 | 85.71 | 80.95 | 78.95 | 84.21 | 82.01 | |
| RS-AgentLLM Backbone=gpt-3.5-turbo-11062024.06 | 94.74 | 84.21 | 95.24 | 57.14 | 85.71 | 78.95 | 100 | 84.89 | |
| RS-ChatGPTLLM Backbone=gpt-3.5-turbo-11062024.06 | 47.37 | 78.95 | 90.48 | 42.86 | 90.48 | 57.89 | 100 | 72.66 |