精选理由
法律 AI 团队终于有了可审计的验证基准——每个任务 50+ 标准像人类一样逐条检查，做法律智能体开发的可以直接参考。
Harvey 推出 LAB 基准，模拟人类验证方式，每个任务包含 50 多条通过标准，每条标准由独立法官调用评估。LangChain Labs 与 Harvey 合作，探索如何在大规模场景下提升效率。该基准可审计性强，但成本较高，合作旨在优化这一过程。
原文 · LangChain
.@Harvey’s LAB benchmark approaches verification like a human would. Every task in a dataset has cr...
. @Harvey ’s LAB benchmark approaches verification like a human would. Every task in a dataset has criteria for the task to pass. Legal agents can have 50+, with each one having its own judge call. It’s easy to audit, but can be expensive at scale. LangChain Labs teamed up with Harvey to see how we can make this more efficient. 💬 2 🔄 1 ❤️ 7 👀 678 📊 3 ⚡