List your company

Multihop Question Answering on 2WikiMultiHopQA Out-of-Distribution (Context Scaling)

70.3Accuracy (8K Context)MemReread
32.13242.04151.9561.859May 11, 2026

Evaluation results

MethodLinks
MemRereadScale=4B2026.0570.371.159.464.154.755.546.945.358.4
MemAgentScale=4B2026.056851.643.839.139.839.135.239.844.6
ReMemR1Scale=4B2026.0564.154.744.537.541.449.237.541.446.3
MemRereadScale=1.7B2026.0537.536.737.52528.1——2530.3
ReMemR1Scale=1.7B2026.0535.234.439.121.927.33221.122.729.2
MemAgentScale=1.7B2026.0533.636.724.223.421.121.124.221.925.8