Amazon
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Amazon Sports → Cloth time-aware (test)
4.09HR@5
9
Amazon Home → Elec time-aware (test)
2.44HR@5
9
Amazon Elec → Home time-aware (test)
3.23HR@5
9
Amazon 4DBInfer (test)
0.774AUC
9
Amazon Games
83.74HR@5
9
Amazon2M (10%/10%/80%)
90.4Accuracy
9
Amazon HGB (test)
0.941ROC-AUC
9
Amazon-670K large scale XMC (test)
36.16PSP@1
9
Amazon Toys-Games public (test)
93.09AUC
9
Amazon Home-Kitchen public (test)
0.895AUC
9
Amazon Office-Products public (test)
97.73AUC
9
Amazon Full (test)
36.4Test Error
9
Amazon Toys (test)
18.39HR@5
8
Amazon
10.6Backward Transfer (BWT)
8
Amazon-books
11.67Recall@10
8
Amazon (test)
89.85Accuracy
8
Amazon Instruments (Tail items)
1.17HR@5
8
Amazon Spurious 90% (train)
99.6F1 Score
8
Amazon
1.9Generalization Gap (Δ)
8
Amazon-Comp
87.01Accuracy
8
Amazon-5
65Accuracy
8
Amazon LLM-generated Qwen-DSR1
99.93AUC
8
Amazon LLM-generated Llama3
99.94AUC
8
Amazon LLM-generated Qwen2
99.98AUC
8
Amazon WILDS
54.710th Percentile Accuracy
8