Interactive Decision Making on Virtualhome
63.3Success RateHIPIF
Evaluation Results
| Method | Links | |
|---|---|---|
| HIPIFType=RL2026.06 | 63.3 | |
| GIGPOType=RL2026.06 | 60.9 | |
| HISRType=RL2026.03 | 59.1 | |
| HISR -w/o SPRType=RL2026.03 | 57.9 | |
| HISR -w/o AGSType=RL2026.03 | 57.5 | |
| HISR -w/o HIMType=RL2026.03 | 55.1 | |
| SPAType=RL2026.03 | 53.4 | |
| HISR -w/o BOTHType=RL2026.03 | 53.4 | |
| RFTType=BC2026.03 | 53.1 | |
| DPOType=BC2026.03 | 52.8 | |
| RAGENType=RL2026.03 | 52.1 | |
| SFTType=BC2026.03 | 51.8 | |
| GRPOType=RL2026.03 | 51.2 | |
| PPOType=RL2026.06 | 51.2 | |
| PPOType=RL2026.03 | 51 | |
| Hiagent+GRPOType=RL2026.06 | 48.9 | |
| RLOOType=RL2026.06 | 47.1 | |
| GRPOType=RL2026.06 | 46.9 | |
| HiperType=RL2026.06 | 45.5 | |
| RL-GCDType=RL2026.06 | 42.9 | |
| Gemini2.5proType=PE2026.03 | 31.7 | |
| Gemini2.5proType=PE2026.06 | 31.7 | |
| Gemini2.5pro (Hiagent)Type=PE2026.06 | 28.1 | |
| GPT4o (Hiagent)Type=PE2026.06 | 21.1 | |
| GPT4oType=PE2026.03 | 20.8 | |
| GPT4oType=PE2026.06 | 20.8 | |
| HiagentType=PE2026.06 | 18.8 | |
| ReflexionType=PE2026.06 | 18.2 | |
| ReactType=PE2026.06 | 14.1 | |
| Qwen2.5-3B-InstructType=PE2026.06 | 10.2 | |
| Llama3.2Type=PE2026.03 | 1.2 |