Interactive LLM Agent Task Success on ALFWorld Seen
89.6Success RateQ-Evolve
Evaluation Results
| Method | Links | |
|---|---|---|
| Q-EvolveBackbone=Llama-3-8B-Instruct2026.06 | 89.6 | |
| ETO + MPOBackbone=Llama-3-8B-Instruct2026.06 | 85 | |
| SFT + MPOBackbone=Llama-3-8B-Instruct2026.06 | 80.7 | |
| KnowAgentBackbone=Llama-3-8B-Instruct2026.06 | 80 | |
| SFTBackbone=Llama-3-8B-Instruct2026.06 | 79.3 | |
| WKMBackbone=Llama-3-8B-Instruct2026.06 | 77.5 | |
| ETOBackbone=Llama-3-8B-Instruct2026.06 | 77.1 |