跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2609.16023 · v2

Are We Grading Properly? Understanding Failure Modes in Medical Benchmarks

首次提交 2026-09-10 09:18
版本更新 2026-11-19 04:33
雷达观测 2026-11-21 12:18
arXiv ↗PDF ↗

摘要

Medical evaluation is shifting from static option-based questioning to realistic clinical scenarios with open-ended output modes. Grading these at scale naively, however, is expensive, and rubric-based evaluation has become the dominant scalable alternative. We ask what happens when the rubrics themselves are not airtight, and whether such flaws can be detected and corrected. We apply RIFT, a global rubric failure taxonomy, to three clinical benchmarks (HealthBench Professional, LiveMedBench, and MentalHealthBench), and find failure modes are meaningful: on HealthBench Professional an LLM judge flags 29.6% of criteria as non-atomic and 65.4% as misaligned/rigid, although these rates vary severalfold across five detection judges. Then, we show that these flaws are meaningful and not simply cosmetic. As an example, rewriting bundled criteria of the form "at least one of / all of the following" as equally weighted children and regrading identical responses shifts scores by up to 15.9 percentage points on affected conversations, with disjunctive bundles inflating scores and conjunctive bundles deflating them. Across six answering models and two common graders, these shifts leave the model ranking unchanged, and their net direction depends on how many bundled criteria are penalties. We also find that RIFT can under-detect bundling on clinical rubrics, by an amount that depends on the detection judge: on the same LiveMedBench criteria, judges flag 3.8-15.9% as non-atomic against an estimated 7-19% that are genuinely bundled.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-19 04:33 · 最新观测版本

Are We Grading Properly? Understanding Failure Modes in Medical Benchmarks

Medical evaluation is shifting from static option-based questioning to realistic clinical scenarios with open-ended output modes. Grading these at scale naively, however, is expensive, and rubric-based evaluation has become the dominant scalable alternative. We ask what happens when the rubrics themselves are not airtight, and whether such flaws can be detected and corrected. We apply RIFT, a global rubric failure taxonomy, to three clinical benchmarks (HealthBench Professional, LiveMedBench, and MentalHealthBench), and find failure modes are meaningful: on HealthBench Professional an LLM judge flags 29.6% of criteria as non-atomic and 65.4% as misaligned/rigid, although these rates vary severalfold across five detection judges. Then, we show that these flaws are meaningful and not simply cosmetic. As an example, rewriting bundled criteria of the form "at least one of / all of the following" as equally weighted children and regrading identical responses shifts scores by up to 15.9 percentage points on affected conversations, with disjunctive bundles inflating scores and conjunctive bundles deflating them. Across six answering models and two common graders, these shifts leave the model ranking unchanged, and their net direction depends on how many bundled criteria are penalties. We also find that RIFT can under-detect bundling on clinical rubrics, by an amount that depends on the detection judge: on the same LiveMedBench criteria, judges flag 3.8-15.9% as non-atomic against an estimated 7-19% that are genuinely bundled.

查看此版本解读 →查看此版本原文 ↗