List your company

Multi-hop Question Answering on PopQA

38.36AccuracyQwen2.5-7B-Instruct + MemTrain + MEM1
11.559218.517125.47532.4329Jun 2, 2026

Evaluation results

MethodLinks
Qwen2.5-7B-Instruct + MemTrain + MEM1Backbone=Qwen2.5-7B-Instruct, Memory Tool=MEM1, Training=MemTrain2026.0638.36
Qwen3-4B-Instruct-2507 + MemTrain + MEM1Backbone=Qwen3-4B-Instruct-2507, Memory Tool=MEM1, Training=MemTrain2026.0637.91
Qwen2.5-7B-Instruct + MEM1Backbone=Qwen2.5-7B-Instruct, Memory Tool=MEM12026.0627.13
Qwen3-4B-Instruct-2507 + MEM1Backbone=Qwen3-4B-Instruct-2507, Memory Tool=MEM12026.0623.62
Qwen3-4B-Instruct-2507Backbone=Qwen3-4B-Instruct-25072026.0621.81
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B-Instruct2026.0612.59