Multihop Question Answering on 2WikiMultiHopQA Out-of-Distribution (Context Scaling)
70.3Accuracy (8K Context)MemReread
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| MemRereadScale=4B2026.05 | 70.3 | 71.1 | 59.4 | 64.1 | 54.7 | 55.5 | 46.9 | 45.3 | 58.4 | |
| MemAgentScale=4B2026.05 | 68 | 51.6 | 43.8 | 39.1 | 39.8 | 39.1 | 35.2 | 39.8 | 44.6 | |
| ReMemR1Scale=4B2026.05 | 64.1 | 54.7 | 44.5 | 37.5 | 41.4 | 49.2 | 37.5 | 41.4 | 46.3 | |
| MemRereadScale=1.7B2026.05 | 37.5 | 36.7 | 37.5 | 25 | 28.1 | — | — | 25 | 30.3 | |
| ReMemR1Scale=1.7B2026.05 | 35.2 | 34.4 | 39.1 | 21.9 | 27.3 | 32 | 21.1 | 22.7 | 29.2 | |
| MemAgentScale=1.7B2026.05 | 33.6 | 36.7 | 24.2 | 23.4 | 21.1 | 21.1 | 24.2 | 21.9 | 25.8 |