Towards a Deterministic Math Solver for Clinical Language Models

  • 类型:arxiv
  • 标识:2609.10728
  • 链接:https://arxiv.org/abs/2609.10728
  • 主分类:evaluation
  • 形态:method
  • TLDR:Large language models are unreliable at arithmetic, which is a problem for clinical calculators where a single numerical error changes the recommendation. The standard response is to hardcode each calculator as a validated function, one at a time. We test an alternative: the model does not calculate. Instead, it writes case-specific Python that a restricted local executor runs as a deterministic solver, and the model's task reduces to deciding how to use it. We evaluate this Program-Solve interface on MedCalc-Bench Verified (1,100 cases, 55 calculators) against direct model arithmetic and a ha
  • 待LLM分类:否
  • 来源文件
  • /inbox/tom/_candidates/2026-09-15-rag-retrieval-reranking-candidates.json
  • /inbox/tom/_candidates/2026-09-15-agent-rag-longcontext-candidates.json