SFGA：一种用于可信SFT数据采购的具有裁决升级的统计优先门控架构
SFGA: A Statistics-First Gating Architecture with Adjudicative Escalation for Trustworthy SFT Data Procurement
- DGrid AI（DGrid人工智能）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
研究如何采购监督微调数据，提出统计优先门控架构SFGA，将采购视为成本感知路由问题，经实验在准确率、F1值和成本上取得良好平衡，还报告辩论路径负面诊断，为测量和校准构建合成基准。
AI中文摘要：
采购监督微调（SFT）数据要求买家在任何下游训练之前决定候选语料库是否值得获取。我们提出了SFGA，这是一种统计优先的门控架构，将采购视为基于多样性、效用和冗余三个内在质量轴的成本感知路由问题。廉价的盲目测量被总结为带有置信区间的每个轴的估计值；只有当区间紧密、样本量足够且轴一致时，门才接受决策，否则将情况升级到支持购买和支持拒绝的法官之间的裁决辩论，由主审裁决解决。在12个数据集的受控基准测试（三个轴上的2×3×2网格）上，有5个种子，门在每单位0.017美元的情况下达到0.90的准确率和0.83的F1值，介于始终验证基线（0.75）和预言机上限（0.98）之间，同时花费低于始终升级（0.020美元）。我们进一步报告了辩论路径的诚实负面诊断：反方胜率为0.80（p≈3×10−6），在倡导者交换下52%的立场翻转率暴露了天真的大语言模型法官会隐藏的负面和位置偏差。我们将注入旋钮评估明确地构建为用于测量保真度和路由校准的受控合成基准，并将外部有效性界定为未来的工作。
英文摘要：
Procuring supervised fine-tuning (SFT) data forces a buyer to decide, before any downstream training, whether a candidate corpus is worth acquiring. We present \sys{}, a statistics-first gating architecture that treats procurement as a cost-aware routing problem over three intrinsic quality axes -- diversity, utility, and redundancy. Cheap blind measurements are summarised into per-axis estimates with confidence intervals; a gate accepts a decision only when intervals are tight, sample sizes are adequate, and the axes agree, otherwise it escalates the case to an adjudicative debate between a buy-advocate and a reject-advocate judge, resolved by a presiding verdict. On a controlled benchmark of 12 datasets ($2{\times}3{\times}2$ grid over the three axes) with 5 seeds, the gate reaches 0.90 accuracy and 0.83 $F_1$ at \$0.017 per unit, sitting between an always-verify baseline (0.75) and an oracle upper bound (0.98) while spending less than always-escalate (\$0.020). We further report honest negative diagnostics of the debate path: a con-side win rate of 0.80 ($p\approx3{\times}10^{-6}$) and a 52\% position-flip rate under advocate swapping expose negativity and positional biases that a naive LLM-judge would hide. We frame the injected-knob evaluation explicitly as a controlled synthetic benchmark for measurement fidelity and routing calibration, and delimit external validity as future work.
