研究质疑工具图记忆对多轮工具调用的实际作用
Rethinking the Role of Tool Graph Memory in Multi-Turn Tool Use
这篇论文做了 9 个模型 2 个基准的实验，发现流行的工具图记忆基本没用，换文本存经验反而更好还省钱。
一项 arXiv 论文在 9 个模型和 2 个基准上评估了工具图记忆（tool graph memory）。结果显示它相对基线很少带来显著提升，且有效果时 LLM 调用成本超过 2 倍。其引导经常是错的，随机或无引导也能保留大部分收益，说明增益来自额外计算而非记忆本身。把同样的经验存成文本洞察，几乎零额外 LLM 调用就超过了工具图记忆。
Rethinking the Role of Tool Graph Memory in Multi-Turn Tool Use
Large language model (LLM) agents solve complex tasks by calling external tools over multiple turns. Recent work has proposed tool graph memory, which stores past experience as tool transitions to guide subsequent selection. However, it remains unclear whether this component consistently improves agents' tool use or undermines it. We evaluate tool graph memory on nine models and two benchmarks, finding that it seldom improves significantly over baselines, and that its gains, where present, cost over twice the LLM calls. The memory's guidance is often incorrect, and random or no guidance retains most of the gain, implying that the benefit stems from extra computation rather than the memory itself. Moreover, even with the correct next tool suggested, many failures remain, such as giving up without a tool call. Storing the same experience instead as textual insights outperforms tool graph memory with almost no extra LLM calls. This suggests that memory serves better as a soft prior than as a prescription of the next tool, leading to a more effective and generalizable solution.
