实测个人智能体记忆并非越多越好:10 行左右最佳
做个人智能体的可以看看:Haiku 4.5 实测记忆超过 10 行反而更糟,计数类任务该写代码而不是写记忆规则。
实验用 Claude Haiku 4.5 测试记忆条数与规则违规率的关系:无记忆时违规率 77%,10 行记忆降到 20%,更长后回升到约 25%。
需要计数的任务(如累计消费总额)用文字规则会失败 44%,改用代码记录则失败 0%。靠智能体自己改写记忆的方法最好也只能到 48% 违规率,而直接告知全部偏好可低至 7.
1%。
More memory does not keep helping personal agents, because relevant notes help up to a point and then extra lines start making the agent miss rules.
A personal agent can't learn every user preference through memory notes, and more notes eventually make it worse, so use code for anything it must count or track and keep memory short.
Written rules work for style, like signing texts with the user's first name. For a running spending total, a stated rule still failed 44% of the time, while code that kept the total failed 0%.
Memory size has a sweet spot. With Claude Haiku 4.5, violations dropped from 77% with no memory to 20% at 10 lines, then rose to about 25% with longer memories.
Agents that rewrite their own memory from user complaints improve early, then stall. The best methods ended near 48% violations, against 7.1% when the agent was simply told every preference.
本条情报归属 AITOP「技巧」核心栏目,重点聚焦北美前沿实验室与硅谷科技巨头。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:Claude Haiku 4.5、智能体、记忆管理、提示词工程,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「rohanpaul_ai」(@rohanpaul_ai),经 AITOP 多维评分管线结构化提炼。截至目前,全网已有 2 家独立信源对该事件进行交叉印证与跟进报道(含 arXiv cs.AI 等),属于近期高关注度行业信号。
- arXiv cs.AI10-02 16:55原文