Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Long-context Memory Retrieval and Reasoning on LoCoMo 32K

46.39F1 Score

MemSifter

26.671631.790836.9142.0292Mar 3, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.03
46.39
2026.03
41.94
2026.03
41.79
2026.03
41.04
2026.03
40.96
2026.03
39.81
2026.03
39.19
2026.03
38.73
2026.03
36.25
2026.03
36.2
2026.03
35.87
2026.03
33.04
2026.03
31.94
2026.03
31.82
2026.03
31.62
2026.03
30.97
2026.03
29.9
2026.03
29.49
2026.03
28.11
2026.03
27.43