新闻
Patient-specific rubric-based evaluation of large language models across longitudinal cancer care
arXiv cs.AI · 发布于 · 阅读约3分钟
30秒读懂
- 发生了什么
- Researchers created NPC-RubricBench, a benchmark evaluating 30 LLMs across six sequential cancer care phases using 941 patient cases with phase-specific criteria.
- 为何重要
- Matters for engineers building clinical AI systems that must maintain reliability across multiple treatment stages, not just single diagnostic moments.
- 注意
- Strong diagnostic performance did not translate to downstream care quality. Detailed prompts improved decision quality but reduced evidence faithfulness and safety adherence.
- llm
- language model
- reasoning
- eval
- benchmark
这条新闻背后的模式
每个模式都讲清楚技术如何运作、何时值得投入，以及在哪里会失效。
The Agent Architect
每周一个模式、一个权衡、一个生产事故案例。为构建智能体系统的人准备的每周简报。
每周一封邮件,一键退订。您的地址仅用于发送简报。