Embodied Task Completion on ALFWorld (Seen tasks)
97.5Pick Success RateEvolving-RL
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Evolving-RLSkill Injection=true2026.05 | 97.5 | 100 | 100 | 97.4 | 98.6 | |
| GRPOSkill Injection=false2026.05 | 96.7 | 100 | 100 | 90.3 | 96.2 | |
| GRPOSkill Injection=true2026.05 | 95.8 | 98.3 | 98.1 | 96.1 | 97 | |
| Evolving-RLSkill Injection=false2026.05 | 94.2 | 99.1 | 100 | 96.8 | 97.4 | |
| SkillRLSkill Injection=false2026.05 | 93.3 | 72.2 | 80 | 95.2 | 86.2 | |
| MementoSkill Injection=false2026.05 | 79.2 | 39.1 | 76.2 | 64.5 | 64.6 | |
| Base ModelSkill Injection=false2026.05 | 75 | 13 | 60 | 57.4 | 51.9 | |
| ReasoningBankSkill Injection=false2026.05 | 75 | 21.7 | 47.6 | 45.2 | 47.5 | |
| Base ModelSkill Injection=true2026.05 | 74.2 | 28.7 | 56.2 | 45.8 | 50.9 | |
| ExpeLSkill Injection=false2026.05 | 21 | 52 | 71 | 55 | 49.5 |