Long-horizon Question Answering on LoCoMo Overall All Categories 1.0
4.63EM RankMemoryBank
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MemoryBankBackbone=Llama3.2-3B2026.01 | 4.63 | 2.75 | |
| MemoryBankBackbone=Qwen2.5-1.5B2026.01 | 4.63 | 3.75 | |
| MemoryBankBackbone=GPT-4o-mini2026.01 | 4.25 | 4.25 | |
| ReadAgentBackbone=GPT-4o-mini2026.01 | 4.25 | 4.75 | |
| ReadAgentBackbone=Qwen2.5-1.5B2026.01 | 3.88 | 4.5 | |
| ReadAgentBackbone=Llama3.2-3B2026.01 | 3.63 | 4.75 | |
| LoCoMoBackbone=GPT-4o-mini2026.01 | 3.5 | 2 | |
| LoCoMoBackbone=Llama3.2-3B2026.01 | 3.25 | 3.25 | |
| LoCoMoBackbone=Qwen2.5-1.5B2026.01 | 3.25 | 3.25 | |
| MemoryBankBackbone=Llama3.2-1B2026.01 | 3 | 3.75 | |
| A-MemBackbone=Llama3.2-1B2026.01 | 3 | 2 | |
| ReadAgentBackbone=Llama3.2-1B2026.01 | 2.75 | 4.75 | |
| A-MemBackbone=Llama3.2-3B2026.01 | 2.5 | 2.5 | |
| A-MemBackbone=Qwen2.5-1.5B2026.01 | 2.25 | 2.5 | |
| A-MemBackbone=GPT-4o-mini2026.01 | 2 | 2.5 | |
| LoCoMoBackbone=Llama3.2-1B2026.01 | 1.75 | 3 | |
| MemWeaverBackbone=Llama3.2-1B2026.01 | 1.25 | 1.5 | |
| MemWeaverBackbone=GPT-4o-mini2026.01 | 1 | 1.5 | |
| MemWeaverBackbone=Llama3.2-3B2026.01 | 1 | 1.75 | |
| MemWeaverBackbone=Qwen2.5-1.5B2026.01 | 1 | 1 |