Question Answering on RULER QA
75AccuracyMemReread
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MemRereadScale=4B, Context Length=16K2026.05 | 75 | — | |
| ReMemR1Scale=4B, Context Length=8K2026.05 | 71.9 | — | |
| ReMemR1Scale=4B, Context Length=16K2026.05 | 70.1 | — | |
| MemAgentScale=4B, Context Length=16K2026.05 | 69.5 | — | |
| MemAgentScale=4B, Context Length=8K2026.05 | 67.2 | — | |
| MemRereadScale=4B, Context Length=8K2026.05 | 67.2 | — | |
| ReMemR1Scale=4B, Context Length=32K2026.05 | 63.5 | — | |
| MemRereadScale=4B, Context Length=64K2026.05 | 63.3 | — | |
| MemRereadScale=4B, Context Length=128K2026.05 | 63.3 | — | |
| ReMemR1Scale=4B, Context Length=64K2026.05 | 62.7 | — | |
| ReMemR1Scale=4B, Context Length=256K2026.05 | 61.7 | — | |
| MemRereadScale=4B, Context Length=256K2026.05 | 61.6 | — | |
| MemRereadScale=4B, Context Length=32K2026.05 | 60.9 | — | |
| MemRereadScale=4B, Context Length=1M2026.05 | 60.4 | — | |
| ReMemR1Scale=4B, Context Length=128K2026.05 | 60.2 | — | |
| MemRereadScale=4B, Context Length=512K2026.05 | 59.7 | — | |
| MemAgentScale=4B, Context Length=1M2026.05 | 57 | — | |
| ReMemR1Scale=4B, Context Length=512K2026.05 | 56.3 | — | |
| ReMemR1Scale=4B, Context Length=1M2026.05 | 54.2 | — | |
| MemAgentScale=4B, Context Length=32K2026.05 | 53.1 | — | |
| MemAgentScale=4B, Context Length=64K2026.05 | 51.6 | — | |
| MemAgentScale=4B, Context Length=128K2026.05 | 51.6 | — | |
| MemAgentScale=4B, Context Length=256K2026.05 | 50.8 | — | |
| MemAgentScale=4B, Context Length=512K2026.05 | 45.3 | — | |
| MemAgentScale=4B2026.05 | — | 55.8 | |
| MemRereadScale=4B2026.05 | — | 63.9 | |
| ReMemR1Scale=4B2026.05 | — | 62.6 |