Question Answering on HotpotQA distractor setting
67.8F1 ScoreHAGE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HAGELLM Backbone=GPT-4o-mini2026.05 | 67.8 | 82.4 | |
| MAGMALLM Backbone=GPT-4o-mini2026.05 | 64 | 80.7 | |
| MemSkillLLM Backbone=GPT-4o-mini2026.05 | 57.9 | 77.9 | |
| MemoryOSLLM Backbone=GPT-4o-mini2026.05 | 47.7 | 59.2 | |
| A-MEMLLM Backbone=GPT-4o-mini2026.05 | 43.3 | 54.7 | |
| HAGELLM Backbone=Qwen2.5-3B2026.05 | 42.9 | 52.7 | |
| MemoryOSLLM Backbone=Qwen2.5-3B2026.05 | 35 | 45.9 | |
| MAGMALLM Backbone=Qwen2.5-3B2026.05 | 33.7 | 42.4 | |
| A-MEMLLM Backbone=Qwen2.5-3B2026.05 | 18.6 | 41.6 | |
| MemSkillLLM Backbone=Qwen2.5-3B2026.05 | 17.9 | 24.7 | |
| NemoriLLM Backbone=GPT-4o-mini2026.05 | 13.1 | 62.4 | |
| NemoriLLM Backbone=Qwen2.5-3B2026.05 | 9.1 | 33.2 |