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

Factual Accuracy and Reasoning on LoCoMo

88.53Single-hop Accuracy

Full-Context (Ceiling)

35.770849.467963.16576.8621Mar 20, 2026Mar 31, 2026Apr 11, 2026Apr 22, 2026May 3, 2026May 14, 2026May 25, 2026
Updated 9d ago

Evaluation Results

MethodLinks
2026.03
88.5377.771.8892.787.52
2026.03
87.8772.763.5480.3781.95
2026.03
79.4369.1673.9683.3379.09
2026.03
74.4761.0667.7186.9278.05
2026.03
62.4157.3244.7966.4762.47
2026.05
41.832.117.520.234
2026.05
41.628.915.322.933.7
2026.05
40.330.116.818.832.5
2026.05
37.823.917.918.630