Long-context Question Answering on LoCoMo (accuracy)
66.5Accuracy (LoCoMo QA)Mem-T
Evaluation results
| Method | Links | |
|---|---|---|
| Mem-TLLM judge=true2026.06 | 66.5 | |
| LightMemLLM judge=true2026.06 | 62 | |
| Long contextLLM judge=true2026.06 | 61.5 | |
| HippoRAGLLM judge=true2026.06 | 58 | |
| SimpleMemLLM judge=true2026.06 | 56.5 | |
| AMA-AgentLLM judge=true2026.06 | 54.5 | |
| PlugMemLLM judge=true2026.06 | 53 | |
| MemRLLLM judge=true2026.06 | 47.5 | |
| DCI-Lite+SumLLM judge=true2026.06 | 46.5 | |
| DCI-LiteLLM judge=true2026.06 | 45 |