AlfWorld
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
ALFWorld Large context window
100Pick Success Rate
8
ALFWorld Small context window
95.8Pick Success Rate
8
ALFWorld
93.6Success Rate
8
ALFWorld (OOD)
97.25Accuracy
8
AlfWorld (out of distribution)
58.95Accuracy
8
ALFWorld (qualitative context)
99Success Rate (SR)
8
ALFWorld (In-domain)
91.4Success Rate
8
ALFWorld Text N=134 unseen tasks Misaligned
501,049Token Count
7
ALFWorld (out-of-distribution)
86.3Out-Success Rate
7
ALFWorld-B
22.89Cumulative Return
7
ALFWorld Unseen
90.3Success Rate
7
ALFWorld Seen
89.6Success Rate
7
ALFWorld (test)
94.2Pick & Place (P&P)
7
ALFWorld
11.8Steps
7
ALFWorld (Mixed)
98.6In-Domain Completion Rate
7
AlfWorld
59Average Reward
7
ALFWorld OOD v1
0.994SR
7
ALFWorld standard evaluation set (134 tasks)
97Pass@1 Accuracy
7
ALFWorld tool-use (test)
77.8Success Rate
7
ALFWorld unseen
99.3Success Rate (SR)
6
ALFWorld
99.25Success Rate
6
ALFWorld (iter3)
0.8Rho
6
ALFWorld (iter1)
0.82Rho
6
ALFWorld
94.6Max Accuracy
6
ALFWorld
30.7Accuracy
6