Multi-hop Question Answering on HotpotQA 400 documents
61.46F1 ScoreE-mem
Evaluation Results
| Method | Links | |
|---|---|---|
| E-memBackbone=GPT4o-mini2026.01 | 61.46 | |
| E-memBackbone=Qwen2.5-14B2026.01 | 61.13 | |
| Long-ContextBackbone=GPT4o-mini2026.01 | 56.56 | |
| GAMBackbone=GPT4o-mini2026.01 | 54.75 | |
| RAGBackbone=GPT4o-mini2026.01 | 52.71 | |
| RAGBackbone=Qwen2.5-14B2026.01 | 51.81 | |
| Long-ContextBackbone=Qwen2.5-14B2026.01 | 49.75 | |
| GAMBackbone=Qwen2.5-14B2026.01 | 48.4 | |
| LIGHTMEMBackbone=GPT4o-mini2026.01 | 40.93 | |
| LIGHTMEMBackbone=Qwen2.5-14B2026.01 | 37.3 | |
| A-memBackbone=GPT4o-mini2026.01 | 33.9 | |
| Mem0Backbone=GPT4o-mini2026.01 | 32.85 | |
| Mem0Backbone=Qwen2.5-14B2026.01 | 30.12 | |
| A-memBackbone=Qwen2.5-14B2026.01 | 27.04 | |
| MEMORYOSBackbone=GPT4o-mini2026.01 | 26.47 | |
| MEMORYOSBackbone=Qwen2.5-14B2026.01 | 24.58 |