% Author: Simon-Pierre Boucher — simon-pierre.boucher@uqo.ca % generated by scripts/20_evaluate.py — do not edit \begin{tabular}{lrrrrrr} \toprule Arm & {Cost (USD)} & {Latency (s)} & {Tokens} & {Total cost (USD)} & {Mean $|e|$} & {N} \\ \midrule A (hedonic) & \num{0.0000} & \num{0.0} & \num{0} & \num{0.00} & \num{0.198} & \num{83489} \\ B (hierarchical) & \num{0.0000} & \num{0.0} & \num{0} & \num{0.00} & \num{0.185} & \num{83489} \\ C (catboost) & \num{0.0000} & \num{0.0} & \num{0} & \num{0.00} & \num{0.191} & \num{83489} \\ C (gbm) & \num{0.0000} & \num{0.0} & \num{0} & \num{0.00} & \num{0.193} & \num{83489} \\ C (xgboost) & \num{0.0000} & \num{0.0} & \num{0} & \num{0.00} & \num{0.192} & \num{83489} \\ D (avm) & \num{0.0000} & \num{0.0} & \num{0} & \num{0.00} & \num{0.180} & \num{83489} \\ \midrule E (single llm) & \num{0.0228} & \num{17.6} & \num{7343} & \num{1.37} & \num{0.195} & \num{60} \\ F (independent) & \num{0.2419} & \num{81.2} & \num{77650} & \num{14.51} & \num{0.186} & \num{60} \\ G (debate r2) & \num{1.5783} & \num{341.6} & \num{537705} & \num{94.70} & \num{0.189} & \num{60} \\ \bottomrule \end{tabular}