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

Tool Calling on HotpotQA (evaluation)

70.4Accuracy

Baseline (normal SFT)

11.6426.89542.1557.405May 13, 2026
Updated 20d ago

Evaluation Results

MethodLinks
2026.05
70.42.3
2026.05
68.71.1
2026.05
68.62.7
2026.05
67.51.2
2026.05
33.4-
2026.05
23.6-
2026.05
20.1-
2026.05
17.2-
2026.05
15.1-
2026.05
13.9-