关于指标的疑问 || Questions about indicators

[effort-yq]

想请问下，这个locomo的得分78.70%是作为Openclaw插件测试得出的呢还是作为普通测记忆系统的流程得出的呢？？
🎯 更准：[准确率提升 48.77%] 在 LOCOMO 基准测试中，相比于 full-context 更准确（78.70% VS 52.9%）
⚡ 更快：[响应速度快 91.83%] 相比于 full-context，检索的 p95 延迟显著降低（1.44s VS 17.12s）
💰 更省：[Token 用量降低 96.53%] 相比于full-context，在不牺牲性能的前提下显著降低成本（0.9k VS 26k）

---

I would like to ask, is this locomo score 78.70% obtained as an Openclaw plug-in test or as a normal test memory system process? ?
🎯 More accurate: [Accuracy increased by 48.77%] In the LOCOMO benchmark, it is more accurate than full-context (78.70% VS 52.9%)
⚡ Faster: [91.83% faster response] Significantly lower p95 latency for retrieval compared to full-context (1.44s VS 17.12s)
💰 More savings: [Token usage reduced by 96.53%] Compared with full-context, the cost is significantly reduced without sacrificing performance (0.9k VS 26k)

[Teingi]

locomo 的得分是基于locomo 数据集的压测结果的LLM评分得出来的，这也是市面上的memory产品大部分采用的评测集。非Openclaw插件的场景测试的。Openclaw插件 支持替换Openclaw的原来的记忆组件的功能，算是功能上的平替。

---

The score of locomo is based on the LLM score of the stress test results of the locomo data set, which is also the evaluation set used by most memory products on the market. Tested in non-Openclaw plug-in scenarios. The Openclaw plug-in supports the function of replacing the original memory component of Openclaw, which can be regarded as a functional replacement.

[effort-yq]

谢谢您的答复

---

Thank you for your reply