List your company

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

24.4EMESearch-base
7.44811.84916.2520.651Oct 1, 2025

Evaluation results

MethodLinks
ESearch-baseModel Scale=7B2025.1024.437.1
ESearch-instructModel Scale=7B2025.1024.135.8
ESearch-baseModel Scale=3B2025.1023.634.5
ESearch-instructModel Scale=3B2025.1023.233.9
StepSearch-instructModel Scale=7B2025.1022.631.2
ZeroSearch-instructModel Scale=7B2025.1021.932
StepSearch-baseModel Scale=7B2025.1021.632.4
Search-R1-baseModel Scale=7B2025.1020.629
ZeroSearch-baseModel Scale=7B2025.1020.128.6
Research-instructModel Scale=7B2025.1018.427.1
StepSearch-baseModel Scale=3B2025.1018.127.3
Search-R1-instructModel Scale=7B2025.1018.126.2
R-Search-instruct-GRPOModel Scale=7B2025.1017.926
StepSearch-instructModel Scale=3B2025.1017.426.1
SSRL-instructModel Scale=7B2025.1015.323.8
Research-baseModel Scale=7B2025.1014.323
R-Search-instruct-PPOModel Scale=7B2025.1013.320.9
R-Search-instruct-GRPOModel Scale=3B2025.1013.120.8
R-Search-instruct-PPOModel Scale=3B2025.1012.418.7
Search-R1-instructModel Scale=3B2025.101218.8
ZeroSearch-baseModel Scale=3B2025.109.616.4
SSRL-instructModel Scale=3B2025.109.315.6
ZeroSearch-instructModel Scale=3B2025.108.814.5
Search-R1-baseModel Scale=3B2025.108.114.6