\begin{table}[htbp]\centering\footnotesize
\caption{Audit of the rule-based category labels (500-listing stratified sample)}\label{tab:cat_audit}
\begin{adjustbox}{max width=\textwidth}
\begin{tabular}{lrr}\toprule
Rule label & Listings with annotator consensus & Precision of the rule label \\\midrule
Maple \& honey & 24 & 0.50 \\
Fine grocery & 24 & 0.42 \\
Coffee \& tea & 23 & 0.35 \\
Chocolate & 24 & 0.62 \\
Bakery & 25 & 0.40 \\
Meat \& fish & 22 & 0.32 \\
Cheese \& dairy & 24 & 0.58 \\
Beverages & 24 & 0.17 \\
Alcohol & 25 & 0.28 \\
Beauty \& care & 24 & 0.62 \\
Home \& decor & 22 & 0.36 \\
Fashion & 25 & 0.72 \\
Jewellery & 25 & 0.80 \\
Art \& craft & 25 & 0.52 \\
Children & 21 & 0.62 \\
Pets & 24 & 0.29 \\
Outdoor \& sport & 19 & 0.32 \\
Garden & 25 & 0.84 \\
Industrial & 24 & 0.42 \\
Other & 25 & 0.00 \\
\textbf{All} & 474 & \textbf{0.46} \\
\bottomrule\end{tabular}
\end{adjustbox}
\begin{minipage}{0.95\textwidth}\vspace{2pt}\footnotesize\textit{Notes:} 25 random listings per category annotated independently by two LLMs (claude-opus-5-5, gpt-5.5) with the platform's taxonomy; agreement rule vs A 0.45, rule vs B 0.45, A vs B 0.95; consensus exists for 95\% of listings. Precision = share of consensus labels equal to the rule label. Disagreements listed in the replication package.\end{minipage}
\end{table}