arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.01034cs.CLstat.ME
收藏

有限校准机制图:LLM评审团面板

A Finite-Calibration Regime Map for LLM Judge Panels

School of Computer Science and Engineering(计算机科学与工程学院)

Bin Zhu, Yi Xie, Yanghui Rao

AI总结:

研究在有限人工标注预算下,低维堆叠器与联合输出表对LLM评审团面板的校准权衡,提出有限校准面板选择方法,实验表明多数评审输出可加或冗余。

摘要:

我们研究了在有限人工标注预算下,LLM评审团面板应何时使用低维堆叠器与联合输出表进行校准。低维堆叠器估计成本小但忽略交互,而联合表校准器可表示交互但需为单元格计数和未见模式付出代价。我们将此权衡构建为有限校准机制图,并实例化为有限校准面板选择——一种可部署的验证选择器,涵盖评审路径、前缀大小和聚合器家族,并辅以表格和参数估计诊断。在RewardBench、LLMBar、SummEval和Arena100K上,使用包含DeepSeek V4 Flash的七评审池,标量/可靠性聚合在20个真实数据集-预算单元中赢得16个,表明当前评审输出通常是可加或冗余的。受控的校准增长数据显示互补机制:可加标签仍偏好标量,而六路交互选择更大的联合表,其测试MSE从未见质量消失前的0.224降至0.061。因此,实际问题不是“需要多少评审?”,而是下一个评审的信息在可用人工标注下是否可估计。

英文摘要:

Deploying an LLM judge panel spends human labels on fitting a calibrator, constructing candidate judge paths, and validating which candidate to deploy. We study when finite labels should support a low-dimensional stacker or reliability model, and when an unrestricted joint output table is worth its cell-count and unseen-pattern cost. We cast this as a finite-calibration regime map and instantiate it as Finite-Calibration Panel Selection (FCPS), a validation selector over judge path, deployed panel size, and aggregator family with support diagnostics. Across RewardBench, LLMBar, SummEval, and Arena100K with a seven-judge pool, scalar/reliability aggregation has lower MSE than unrestricted joint-table calibration in 16 of 20 real dataset--budget cells by point estimate, while paired 95% intervals exclude zero in 11 cells; richer backoff/shrinkage tables narrow some gaps while preserving the finite-support bottleneck. Controlled calibration-growth data show the opposite regime: when labels contain a six-way interaction, the selected table grows to the interaction-bearing prefix and its MSE falls from 0.224 to 0.061 once unseen mass vanishes. The practical deployment question is whether the next judge's information is estimable under the available human labels.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑