Dialogue Response Generation on LoCoMo
29.1BLEU-4DialogLM + EventWeave
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DialogLM + EventWeaveDomain Specialization=LoCoMo, Memory Augmentation=EventWeave2025.03 | 29.1 | 43.4 | 66 | 7.8 | |
| GPT-4o + EventWeaveMemory Augmentation=EventWeave2025.03 | 28.4 | 42.6 | 65 | 7.6 | |
| GPT-4o + LifeLongMemMemory Augmentation=LifeLongMem2025.03 | 26.5 | 40.7 | 63 | 7 | |
| GPT-4o + MemWalkerMemory Augmentation=MemWalker2025.03 | 25.7 | 40.1 | 62 | 6.7 | |
| GPT-4o + LongMemMemory Augmentation=LongMem2025.03 | 25.3 | 39.8 | 62 | 6.6 | |
| GPT-4o + ProactiveCoTMemory Augmentation=ProactiveCoT2025.03 | 24.9 | 39.2 | 61 | 6.3 | |
| DialogLMDomain Specialization=LoCoMo2025.03 | 22.8 | 37.5 | 60 | 4.9 | |
| GPT-4o (vanilla)Memory Augmentation=None2025.03 | 21.2 | 35.8 | 58 | 4.3 |