Personalized Tool Orchestration on PEToolBench 50-query (test)
82Tool AccuracyMemToolAgent
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MemToolAgentLLM Backbone=Claude 3.5 Sonnet, R=10, prominence=10^-5, Retrieval Setting=second peak2026.06 | 82 | 82 | |
| A-MEMLLM Backbone=Claude 3.5 Sonnet, Retrieval Strategy=Top-15 embedding-based2026.06 | 70 | 82 | |
| Mem0LLM Backbone=Claude 3.5 Sonnet, Retrieval Strategy=Top-15 embedding-based2026.06 | 70 | 82 |