Dialogue Response Generation on MSC
35.8B-4 ScoreEventWeave
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| EventWeaveBackbone=GPT-o12025.03 | 35.8 | 49.1 | 71 | 8.5 | |
| EventWeaveBackbone=DeepSeek-R12025.03 | 35.2 | 48.6 | 70 | 8.4 | |
| GPT-4o + EventWeaveMemory Augmentation=EventWeave2025.03 | 35 | 48.4 | 70 | 8.3 | |
| EventWeaveBackbone=DeepSeek-V32025.03 | 34.5 | 47.9 | 69 | 8.2 | |
| HAHT + EventWeaveDomain Specialization=MSC, Memory Augmentation=EventWeave2025.03 | 34.2 | 46.8 | 68 | 8 | |
| EventWeaveBackbone=Qwen3-14B2025.03 | 34.2 | 47.6 | 69 | 8.2 | |
| LifeLongMemBackbone=GPT-o12025.03 | 33.9 | 48.2 | 70 | 8 | |
| EventWeaveBackbone=Qwen2.5-14B2025.03 | 33.4 | 47 | 68 | 8 | |
| MemWalkerBackbone=GPT-o12025.03 | 33.4 | 47.9 | 69 | 7.7 | |
| LongMemBackbone=GPT-o12025.03 | 33.3 | 47.7 | 69 | 7.8 | |
| LifeLongMemBackbone=DeepSeek-R12025.03 | 33.3 | 47.7 | 69 | 7.9 | |
| GPT-4o + LifeLongMemMemory Augmentation=LifeLongMem2025.03 | 33.1 | 47.5 | 69 | 7.8 | |
| MemWalkerBackbone=DeepSeek-R12025.03 | 32.8 | 47.4 | 68 | 7.6 | |
| LongMemBackbone=DeepSeek-R12025.03 | 32.7 | 47.2 | 68 | 7.7 | |
| GPT-4o + MemWalkerMemory Augmentation=MemWalker2025.03 | 32.6 | 47.2 | 68 | 7.5 | |
| LifeLongMemBackbone=DeepSeek-V32025.03 | 32.6 | 47 | 68 | 7.7 | |
| GPT-4o + LongMemMemory Augmentation=LongMem2025.03 | 32.5 | 47 | 68 | 7.6 | |
| ProactiveCoTBackbone=GPT-o12025.03 | 32.5 | 47.6 | 68 | 7.3 | |
| LifeLongMemBackbone=Qwen3-14B2025.03 | 32.3 | 46.7 | 68 | 7.7 | |
| MemWalkerBackbone=DeepSeek-V32025.03 | 32.1 | 46.7 | 67 | 7.4 | |
| LongMemBackbone=DeepSeek-V32025.03 | 32 | 46.5 | 67 | 7.5 | |
| ProactiveCoTBackbone=DeepSeek-R12025.03 | 31.9 | 47.1 | 67 | 7.2 | |
| MemWalkerBackbone=Qwen3-14B2025.03 | 31.8 | 46.4 | 67 | 7.4 | |
| GPT-4o + ProactiveCoTMemory Augmentation=ProactiveCoT2025.03 | 31.7 | 46.9 | 67 | 7.1 | |
| LongMemBackbone=Qwen3-14B2025.03 | 31.7 | 46.2 | 67 | 7.5 | |
| LifeLongMemBackbone=Qwen2.5-14B2025.03 | 31.5 | 46.1 | 67 | 7.5 | |
| ProactiveCoTBackbone=DeepSeek-V32025.03 | 31.2 | 46.4 | 66 | 7 | |
| MemWalkerBackbone=Qwen2.5-14B2025.03 | 31 | 45.8 | 66 | 7.2 | |
| LongMemBackbone=Qwen2.5-14B2025.03 | 30.9 | 45.6 | 66 | 7.3 | |
| ProactiveCoTBackbone=Qwen3-14B2025.03 | 30.9 | 46.1 | 66 | 7 | |
| ProactiveCoTBackbone=Qwen2.5-14B2025.03 | 30.2 | 45.5 | 65 | 6.8 | |
| GPT-o1Strategy=vanilla2025.03 | 29.2 | 47 | 65 | 5.4 | |
| DeepSeek-R1Strategy=vanilla2025.03 | 28.7 | 46.5 | 64 | 5.3 | |
| GPT-4o (vanilla)Memory Augmentation=None2025.03 | 28.5 | 46.3 | 64 | 5.2 | |
| DeepSeek-V3Strategy=vanilla2025.03 | 28 | 45.8 | 63 | 5.1 | |
| Qwen3-14BStrategy=vanilla2025.03 | 27.8 | 45.4 | 63 | 5.1 | |
| HAHTDomain Specialization=MSC2025.03 | 27.3 | 43 | 63 | 5 | |
| Qwen2.5-14BStrategy=vanilla2025.03 | 27.1 | 44.8 | 62 | 4.9 |