List your company

Multi-hop Question Answering on Bamboogle offline Wiki-18 (test val)

53.4Exact Match (EM)ESearch-base
15.02424.98734.9544.913Oct 1, 2025

Evaluation results

MethodLinks
ESearch-baseModel Scale=7B2025.1053.465.6
StepSearch-baseModel Scale=7B2025.1046.757.3
ESearch-instructModel Scale=7B2025.1045.861.2
ESearch-instructModel Scale=3B2025.1044.658.7
ZeroSearch-instructModel Scale=7B2025.1043.354
Search-R1-baseModel Scale=7B2025.104354.5
Research-instructModel Scale=7B2025.1042.454.4
ZeroSearch-baseModel Scale=7B2025.1041.753.2
ESearch-baseModel Scale=3B2025.1041.452.9
R-Search-instruct-GRPOModel Scale=7B2025.104051.7
StepSearch-instructModel Scale=7B2025.104053.4
Search-R1-instructModel Scale=7B2025.1038.450.1
R-Search-instruct-PPOModel Scale=7B2025.1038.449.1
Research-baseModel Scale=7B2025.1037.344.9
StepSearch-instructModel Scale=3B2025.1034.445.2
SSRL-instructModel Scale=7B2025.1034.446.6
StepSearch-baseModel Scale=3B2025.1032.841.9
R-Search-instruct-PPOModel Scale=3B2025.102635.5
Search-R1-instructModel Scale=3B2025.102434.4
R-Search-instruct-GRPOModel Scale=3B2025.1022.832.7
SSRL-instructModel Scale=3B2025.1021.628.7
ZeroSearch-instructModel Scale=3B2025.1019.329.9
Search-R1-baseModel Scale=3B2025.1017.627
ZeroSearch-baseModel Scale=3B2025.1016.525.6