Conversational Question Answering on LoCoMo (test)
50.7F1 ScoreFull Context
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Full ContextLLM Backbone=GPT-4.1-MINI2026.05 | 50.7 | 84.8 | |
| RecMemLLM Backbone=GPT-4.1-MINI2026.05 | 46.9 | 81.1 | |
| MemoryOSLLM Backbone=GPT-4O-MINI2026.05 | 45.6 | 63.6 | |
| MemoryOSLLM Backbone=GPT-4.1-MINI2026.05 | 44.5 | 67.6 | |
| AMemLLM Backbone=GPT-4.1-MINI2026.05 | 44 | 69.7 | |
| Mem0LLM Backbone=GPT-4O-MINI2026.05 | 42.9 | 58.6 | |
| Mem0LLM Backbone=GPT-4.1-MINI2026.05 | 42.8 | 62.9 | |
| Full ContextLLM Backbone=GPT-4O-MINI2026.05 | 42.3 | 76.4 | |
| RecMemLLM Backbone=GPT-4O-MINI2026.05 | 38.5 | 72.47 | |
| AMemLLM Backbone=GPT-4O-MINI2026.05 | 36.4 | 60.8 | |
| Naive RagLLM Backbone=GPT-4.1-MINI2026.05 | 33.7 | 54.4 | |
| Naive RagLLM Backbone=GPT-4O-MINI2026.05 | 30.9 | 49.7 |