Update notebooks and experimental results (#12)

* added plots to compare calibration errors * updated plotting notebook * added plots for comparison between prompting schemes
socialfoundations · Sep 4, 2024 · 02f1db5 · 02f1db5
1 parent e78de48
commit 02f1db5
Show file tree

Hide file tree

Showing 19 changed files with 1,079 additions and 651 deletions.
diff --git a/notebooks/paper-plots-and-tables.ipynb b/notebooks/paper-plots-and-tables.ipynb
diff --git a/results/imgs/ECE_calibration_comparison_diff_prompts.any-models.numeric-prompt.pdf b/results/imgs/ECE_calibration_comparison_diff_prompts.any-models.numeric-prompt.pdf
diff --git a/results/imgs/ECE_calibration_comparison_diff_prompts.base-models.numeric-prompt.pdf b/results/imgs/ECE_calibration_comparison_diff_prompts.base-models.numeric-prompt.pdf
diff --git a/results/imgs/ECE_calibration_comparison_diff_prompts.ece.any-models.numeric-prompt.pdf b/results/imgs/ECE_calibration_comparison_diff_prompts.ece.any-models.numeric-prompt.pdf
diff --git a/results/imgs/ECE_calibration_comparison_diff_prompts.ece.it-models.numeric-prompt.pdf b/results/imgs/ECE_calibration_comparison_diff_prompts.ece.it-models.numeric-prompt.pdf
diff --git a/results/imgs/ECE_calibration_comparison_diff_prompts.it-models.numeric-prompt.pdf b/results/imgs/ECE_calibration_comparison_diff_prompts.it-models.numeric-prompt.pdf
diff --git a/results/imgs/ECE_calibration_comparison_diff_prompts.roc_auc.any-models.numeric-prompt.pdf b/results/imgs/ECE_calibration_comparison_diff_prompts.roc_auc.any-models.numeric-prompt.pdf
diff --git a/results/imgs/ECE_calibration_comparison_diff_prompts.roc_auc.base-models.numeric-prompt.pdf b/results/imgs/ECE_calibration_comparison_diff_prompts.roc_auc.base-models.numeric-prompt.pdf
diff --git a/results/imgs/ECE_calibration_comparison_diff_prompts.roc_auc.it-models.numeric-prompt.pdf b/results/imgs/ECE_calibration_comparison_diff_prompts.roc_auc.it-models.numeric-prompt.pdf
diff --git a/...lts/tables/ACSEmployment-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=False.tex b/...lts/tables/ACSEmployment-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=False.tex
@@ -0,0 +1,28 @@
+\begin{tabular}{lllllllll}
+\toprule
+ & ece (mult. choice) & brier score loss (mult. choice) & roc auc (mult. choice) & accuracy (mult. choice) & ece & brier score loss & roc auc & accuracy \\
+Model &  &  &  &  &  &  &  &  \\
+\midrule
+GPT 4o mini (it) & 0.28 & 0.29 & 0.79 & 0.65 & 0.23 & 0.23 & 0.80 & 0.73 \\
+Llama 3 70B (it) & 0.17 & \cellcolor{cyan!17.1} 0.19 & \cellcolor{cyan!23.3} 0.85 & \cellcolor{cyan!15.4} 0.73 & \cellcolor{cyan!0.7} 0.05 & \cellcolor{cyan!25.0} 0.14 & \cellcolor{cyan!25.0} 0.88 & \cellcolor{cyan!25.0} 0.81 \\
+Llama 3 70B & 0.25 & 0.26 & \cellcolor{cyan!2.3} 0.82 & 0.52 & 0.05 & \cellcolor{cyan!16.7} 0.15 & \cellcolor{cyan!13.6} 0.86 & \cellcolor{cyan!5.8} 0.78 \\
+Llama 3 8B (it) & 0.07 & \cellcolor{cyan!25.0} 0.19 & 0.79 & \cellcolor{cyan!25.0} 0.74 & 0.08 & 0.17 & 0.82 & 0.77 \\
+Llama 3 8B & 0.34 & 0.34 & 0.76 & \cellcolor{orange!25.0} 0.45 & 0.15 & 0.23 & 0.75 & \cellcolor{orange!25.0} 0.46 \\
+Mixtral 8x22B (it) & \cellcolor{orange!23.0} 0.38 & \cellcolor{orange!8.0} 0.39 & 0.60 & 0.51 & 0.06 & \cellcolor{cyan!21.9} 0.14 & \cellcolor{cyan!24.4} 0.87 & \cellcolor{cyan!16.5} 0.79 \\
+Mixtral 8x22B & 0.21 & 0.24 & \cellcolor{cyan!25.0} 0.86 & 0.52 & 0.15 & 0.18 & 0.82 & \cellcolor{cyan!20.7} 0.80 \\
+Mixtral 8x7B (it) & 0.22 & 0.24 & \cellcolor{cyan!6.8} 0.82 & \cellcolor{cyan!13.7} 0.73 & 0.07 & \cellcolor{cyan!16.7} 0.15 & \cellcolor{cyan!23.2} 0.87 & \cellcolor{cyan!7.2} 0.78 \\
+Mixtral 8x7B & 0.30 & 0.31 & \cellcolor{cyan!1.1} 0.81 & \cellcolor{orange!24.1} 0.45 & 0.08 & 0.17 & 0.81 & 0.73 \\
+Mistral 7B (it) & \cellcolor{orange!6.6} 0.35 & 0.36 & 0.72 & 0.63 & \cellcolor{cyan!7.4} 0.04 & 0.19 & 0.79 & 0.69 \\
+Mistral 7B & 0.26 & 0.30 & 0.76 & \cellcolor{orange!25.0} 0.45 & 0.14 & 0.19 & 0.80 & \cellcolor{cyan!12.9} 0.79 \\
+Yi 34B (it) & 0.14 & 0.21 & 0.79 & 0.69 & 0.15 & 0.21 & 0.81 & 0.51 \\
+Yi 34B & 0.08 & 0.23 & 0.70 & 0.62 & 0.13 & 0.23 & 0.66 & 0.50 \\
+Gemma 7B (it) & \cellcolor{orange!10.0} 0.36 & 0.38 & 0.59 & 0.58 & \cellcolor{cyan!8.8} 0.04 & 0.22 & 0.71 & 0.60 \\
+Gemma 7B & 0.15 & 0.25 & 0.65 & \cellcolor{orange!4.1} 0.48 & \cellcolor{orange!25.0} 0.35 & \cellcolor{orange!25.0} 0.38 & \cellcolor{orange!1.6} 0.50 & 0.51 \\
+Gemma 2B (it) & \cellcolor{orange!25.0} 0.38 & \cellcolor{orange!25.0} 0.41 & \cellcolor{orange!25.0} 0.42 & \cellcolor{orange!16.3} 0.46 & 0.12 & 0.27 & \cellcolor{orange!25.0} 0.46 & \cellcolor{orange!25.0} 0.46 \\
+Gemma 2B & \cellcolor{cyan!25.0} 0.01 & 0.24 & 0.63 & 0.54 & \cellcolor{cyan!25.0} 0.01 & 0.23 & 0.57 & 0.53 \\
+LR & 0.02 & 0.15 & 0.86 & 0.78 & 0.02 & 0.15 & 0.86 & 0.78 \\
+GBM & 0.00 & 0.12 & 0.91 & 0.83 & 0.00 & 0.12 & 0.91 & 0.83 \\
+XGBoost & 0.00 & 0.12 & 0.91 & 0.83 & 0.00 & 0.12 & 0.91 & 0.83 \\
+\bottomrule
+\end{tabular}
+
diff --git a/results/tables/ACSEmployment-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=True.tex b/results/tables/ACSEmployment-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=True.tex
@@ -0,0 +1,28 @@
+\begin{tabular}{lllllllll}
+\toprule
+ & ece (num) & brier score loss (num) & roc auc (num) & accuracy (num) & ece & brier score loss & roc auc & accuracy \\
+Model &  &  &  &  &  &  &  &  \\
+\midrule
+GPT 4o mini (it) & 0.23 & 0.23 & 0.80 & 0.73 & 0.28 & 0.29 & 0.79 & 0.65 \\
+Llama 3 70B (it) & \cellcolor{cyan!0.7} 0.05 & \cellcolor{cyan!25.0} 0.14 & \cellcolor{cyan!25.0} 0.88 & \cellcolor{cyan!25.0} 0.81 & 0.17 & \cellcolor{cyan!17.1} 0.19 & \cellcolor{cyan!23.3} 0.85 & \cellcolor{cyan!15.4} 0.73 \\
+Llama 3 70B & 0.05 & \cellcolor{cyan!16.7} 0.15 & \cellcolor{cyan!13.6} 0.86 & \cellcolor{cyan!5.8} 0.78 & 0.25 & 0.26 & \cellcolor{cyan!2.3} 0.82 & 0.52 \\
+Llama 3 8B (it) & 0.08 & 0.17 & 0.82 & 0.77 & 0.07 & \cellcolor{cyan!25.0} 0.19 & 0.79 & \cellcolor{cyan!25.0} 0.74 \\
+Llama 3 8B & 0.15 & 0.23 & 0.75 & \cellcolor{orange!25.0} 0.46 & 0.34 & 0.34 & 0.76 & \cellcolor{orange!25.0} 0.45 \\
+Mixtral 8x22B (it) & 0.06 & \cellcolor{cyan!21.9} 0.14 & \cellcolor{cyan!24.4} 0.87 & \cellcolor{cyan!16.5} 0.79 & \cellcolor{orange!23.0} 0.38 & \cellcolor{orange!8.0} 0.39 & 0.60 & 0.51 \\
+Mixtral 8x22B & 0.15 & 0.18 & 0.82 & \cellcolor{cyan!20.7} 0.80 & 0.21 & 0.24 & \cellcolor{cyan!25.0} 0.86 & 0.52 \\
+Mixtral 8x7B (it) & 0.07 & \cellcolor{cyan!16.7} 0.15 & \cellcolor{cyan!23.2} 0.87 & \cellcolor{cyan!7.2} 0.78 & 0.22 & 0.24 & \cellcolor{cyan!6.8} 0.82 & \cellcolor{cyan!13.7} 0.73 \\
+Mixtral 8x7B & 0.08 & 0.17 & 0.81 & 0.73 & 0.30 & 0.31 & \cellcolor{cyan!1.1} 0.81 & \cellcolor{orange!24.1} 0.45 \\
+Mistral 7B (it) & \cellcolor{cyan!7.4} 0.04 & 0.19 & 0.79 & 0.69 & \cellcolor{orange!6.6} 0.35 & 0.36 & 0.72 & 0.63 \\
+Mistral 7B & 0.14 & 0.19 & 0.80 & \cellcolor{cyan!12.9} 0.79 & 0.26 & 0.30 & 0.76 & \cellcolor{orange!25.0} 0.45 \\
+Yi 34B (it) & 0.15 & 0.21 & 0.81 & 0.51 & 0.14 & 0.21 & 0.79 & 0.69 \\
+Yi 34B & 0.13 & 0.23 & 0.66 & 0.50 & 0.08 & 0.23 & 0.70 & 0.62 \\
+Gemma 7B (it) & \cellcolor{cyan!8.8} 0.04 & 0.22 & 0.71 & 0.60 & \cellcolor{orange!10.0} 0.36 & 0.38 & 0.59 & 0.58 \\
+Gemma 7B & \cellcolor{orange!25.0} 0.35 & \cellcolor{orange!25.0} 0.38 & \cellcolor{orange!1.6} 0.50 & 0.51 & 0.15 & 0.25 & 0.65 & \cellcolor{orange!4.1} 0.48 \\
+Gemma 2B (it) & 0.12 & 0.27 & \cellcolor{orange!25.0} 0.46 & \cellcolor{orange!25.0} 0.46 & \cellcolor{orange!25.0} 0.38 & \cellcolor{orange!25.0} 0.41 & \cellcolor{orange!25.0} 0.42 & \cellcolor{orange!16.3} 0.46 \\
+Gemma 2B & \cellcolor{cyan!25.0} 0.01 & 0.23 & 0.57 & 0.53 & \cellcolor{cyan!25.0} 0.01 & 0.24 & 0.63 & 0.54 \\
+LR & 0.02 & 0.15 & 0.86 & 0.78 & 0.02 & 0.15 & 0.86 & 0.78 \\
+GBM & 0.00 & 0.12 & 0.91 & 0.83 & 0.00 & 0.12 & 0.91 & 0.83 \\
+XGBoost & 0.00 & 0.12 & 0.91 & 0.83 & 0.00 & 0.12 & 0.91 & 0.83 \\
+\bottomrule
+\end{tabular}
+
diff --git a/results/tables/ACSIncome-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=False.tex b/results/tables/ACSIncome-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=False.tex
@@ -0,0 +1,28 @@
+\begin{tabular}{lllllllll}
+\toprule
+ & ece (mult. choice) & brier score loss (mult. choice) & roc auc (mult. choice) & accuracy (mult. choice) & ece & brier score loss & roc auc & accuracy \\
+Model &  &  &  &  &  &  &  &  \\
+\midrule
+GPT 4o mini (it) & 0.24 & 0.24 & \cellcolor{cyan!17.6} 0.85 & 0.74 & \cellcolor{cyan!25.0} 0.05 & \cellcolor{cyan!25.0} 0.16 & \cellcolor{cyan!20.6} 0.83 & \cellcolor{cyan!24.4} 0.78 \\
+Llama 3 70B (it) & 0.27 & 0.27 & \cellcolor{cyan!25.0} 0.86 & 0.69 & 0.25 & 0.23 & \cellcolor{cyan!22.1} 0.84 & 0.67 \\
+Llama 3 70B & 0.20 & \cellcolor{cyan!14.9} 0.20 & \cellcolor{cyan!20.8} 0.86 & 0.70 & 0.27 & 0.24 & \cellcolor{cyan!6.7} 0.82 & 0.54 \\
+Llama 3 8B (it) & 0.32 & 0.30 & \cellcolor{cyan!13.3} 0.85 & 0.62 & 0.23 & 0.23 & \cellcolor{cyan!0.1} 0.81 & 0.67 \\
+Llama 3 8B & 0.25 & 0.26 & 0.81 & \cellcolor{orange!20.8} 0.38 & 0.14 & 0.24 & 0.63 & \cellcolor{orange!3.7} 0.40 \\
+Mixtral 8x22B (it) & 0.21 & \cellcolor{cyan!3.6} 0.22 & \cellcolor{cyan!11.2} 0.85 & \cellcolor{cyan!11.1} 0.76 & 0.11 & \cellcolor{cyan!15.5} 0.17 & \cellcolor{cyan!25.0} 0.84 & \cellcolor{cyan!18.9} 0.77 \\
+Mixtral 8x22B & \cellcolor{cyan!13.2} 0.17 & \cellcolor{cyan!21.6} 0.19 & \cellcolor{cyan!16.5} 0.85 & 0.68 & 0.13 & \cellcolor{cyan!3.6} 0.18 & \cellcolor{cyan!9.6} 0.82 & \cellcolor{cyan!2.4} 0.74 \\
+Mixtral 8x7B (it) & \cellcolor{cyan!16.8} 0.16 & \cellcolor{cyan!25.0} 0.18 & \cellcolor{cyan!19.7} 0.86 & \cellcolor{cyan!25.0} 0.78 & 0.10 & \cellcolor{cyan!15.5} 0.17 & \cellcolor{cyan!22.8} 0.84 & \cellcolor{cyan!12.8} 0.76 \\
+Mixtral 8x7B & \cellcolor{cyan!10.6} 0.17 & \cellcolor{cyan!11.5} 0.21 & 0.83 & 0.65 & \cellcolor{cyan!4.0} 0.07 & \cellcolor{cyan!13.1} 0.17 & \cellcolor{cyan!3.7} 0.81 & \cellcolor{cyan!25.0} 0.78 \\
+Mistral 7B (it) & 0.21 & \cellcolor{cyan!4.7} 0.22 & 0.83 & \cellcolor{cyan!16.5} 0.77 & 0.16 & 0.19 & \cellcolor{cyan!14.7} 0.83 & 0.70 \\
+Mistral 7B & 0.20 & 0.23 & 0.80 & 0.73 & \cellcolor{orange!15.7} 0.36 & 0.32 & 0.75 & 0.49 \\
+Yi 34B (it) & \cellcolor{cyan!1.3} 0.19 & \cellcolor{cyan!18.8} 0.19 & \cellcolor{cyan!19.7} 0.86 & 0.72 & 0.22 & 0.21 & 0.80 & 0.48 \\
+Yi 34B & 0.25 & \cellcolor{cyan!2.5} 0.22 & \cellcolor{cyan!13.3} 0.85 & 0.62 & 0.15 & 0.19 & \cellcolor{cyan!17.7} 0.83 & 0.61 \\
+Gemma 7B (it) & \cellcolor{orange!14.7} 0.61 & \cellcolor{orange!4.7} 0.59 & \cellcolor{cyan!3.8} 0.84 & \cellcolor{orange!25.0} 0.37 & 0.33 & 0.30 & 0.78 & 0.42 \\
+Gemma 7B & 0.24 & 0.27 & 0.76 & \cellcolor{orange!25.0} 0.37 & 0.15 & 0.20 & 0.80 & 0.73 \\
+Gemma 2B (it) & \cellcolor{orange!25.0} 0.63 & \cellcolor{orange!25.0} 0.63 & 0.73 & \cellcolor{orange!25.0} 0.37 & 0.28 & 0.31 & \cellcolor{orange!25.0} 0.50 & \cellcolor{orange!25.0} 0.37 \\
+Gemma 2B & \cellcolor{cyan!25.0} 0.14 & 0.25 & \cellcolor{orange!25.0} 0.62 & 0.45 & \cellcolor{orange!25.0} 0.37 & \cellcolor{orange!25.0} 0.37 & \cellcolor{orange!25.0} 0.50 & 0.63 \\
+LR & 0.03 & 0.18 & 0.79 & 0.74 & 0.03 & 0.18 & 0.79 & 0.74 \\
+GBM & 0.01 & 0.13 & 0.89 & 0.81 & 0.01 & 0.13 & 0.89 & 0.81 \\
+XGBoost & 0.00 & 0.13 & 0.90 & 0.82 & 0.00 & 0.13 & 0.90 & 0.82 \\
+\bottomrule
+\end{tabular}
+
diff --git a/results/tables/ACSIncome-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=True.tex b/results/tables/ACSIncome-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=True.tex
@@ -0,0 +1,28 @@
+\begin{tabular}{lllllllll}
+\toprule
+ & ece (num) & brier score loss (num) & roc auc (num) & accuracy (num) & ece & brier score loss & roc auc & accuracy \\
+Model &  &  &  &  &  &  &  &  \\
+\midrule
+GPT 4o mini (it) & \cellcolor{cyan!25.0} 0.05 & \cellcolor{cyan!25.0} 0.16 & \cellcolor{cyan!20.6} 0.83 & \cellcolor{cyan!24.4} 0.78 & 0.24 & 0.24 & \cellcolor{cyan!17.6} 0.85 & 0.74 \\
+Llama 3 70B (it) & 0.25 & 0.23 & \cellcolor{cyan!22.1} 0.84 & 0.67 & 0.27 & 0.27 & \cellcolor{cyan!25.0} 0.86 & 0.69 \\
+Llama 3 70B & 0.27 & 0.24 & \cellcolor{cyan!6.7} 0.82 & 0.54 & 0.20 & \cellcolor{cyan!14.9} 0.20 & \cellcolor{cyan!20.8} 0.86 & 0.70 \\
+Llama 3 8B (it) & 0.23 & 0.23 & \cellcolor{cyan!0.1} 0.81 & 0.67 & 0.32 & 0.30 & \cellcolor{cyan!13.3} 0.85 & 0.62 \\
+Llama 3 8B & 0.14 & 0.24 & 0.63 & \cellcolor{orange!3.7} 0.40 & 0.25 & 0.26 & 0.81 & \cellcolor{orange!20.8} 0.38 \\
+Mixtral 8x22B (it) & 0.11 & \cellcolor{cyan!15.5} 0.17 & \cellcolor{cyan!25.0} 0.84 & \cellcolor{cyan!18.9} 0.77 & 0.21 & \cellcolor{cyan!3.6} 0.22 & \cellcolor{cyan!11.2} 0.85 & \cellcolor{cyan!11.1} 0.76 \\
+Mixtral 8x22B & 0.13 & \cellcolor{cyan!3.6} 0.18 & \cellcolor{cyan!9.6} 0.82 & \cellcolor{cyan!2.4} 0.74 & \cellcolor{cyan!13.2} 0.17 & \cellcolor{cyan!21.6} 0.19 & \cellcolor{cyan!16.5} 0.85 & 0.68 \\
+Mixtral 8x7B (it) & 0.10 & \cellcolor{cyan!15.5} 0.17 & \cellcolor{cyan!22.8} 0.84 & \cellcolor{cyan!12.8} 0.76 & \cellcolor{cyan!16.8} 0.16 & \cellcolor{cyan!25.0} 0.18 & \cellcolor{cyan!19.7} 0.86 & \cellcolor{cyan!25.0} 0.78 \\
+Mixtral 8x7B & \cellcolor{cyan!4.0} 0.07 & \cellcolor{cyan!13.1} 0.17 & \cellcolor{cyan!3.7} 0.81 & \cellcolor{cyan!25.0} 0.78 & \cellcolor{cyan!10.6} 0.17 & \cellcolor{cyan!11.5} 0.21 & 0.83 & 0.65 \\
+Mistral 7B (it) & 0.16 & 0.19 & \cellcolor{cyan!14.7} 0.83 & 0.70 & 0.21 & \cellcolor{cyan!4.7} 0.22 & 0.83 & \cellcolor{cyan!16.5} 0.77 \\
+Mistral 7B & \cellcolor{orange!15.7} 0.36 & 0.32 & 0.75 & 0.49 & 0.20 & 0.23 & 0.80 & 0.73 \\
+Yi 34B (it) & 0.22 & 0.21 & 0.80 & 0.48 & \cellcolor{cyan!1.3} 0.19 & \cellcolor{cyan!18.8} 0.19 & \cellcolor{cyan!19.7} 0.86 & 0.72 \\
+Yi 34B & 0.15 & 0.19 & \cellcolor{cyan!17.7} 0.83 & 0.61 & 0.25 & \cellcolor{cyan!2.5} 0.22 & \cellcolor{cyan!13.3} 0.85 & 0.62 \\
+Gemma 7B (it) & 0.33 & 0.30 & 0.78 & 0.42 & \cellcolor{orange!14.7} 0.61 & \cellcolor{orange!4.7} 0.59 & \cellcolor{cyan!3.8} 0.84 & \cellcolor{orange!25.0} 0.37 \\
+Gemma 7B & 0.15 & 0.20 & 0.80 & 0.73 & 0.24 & 0.27 & 0.76 & \cellcolor{orange!25.0} 0.37 \\
+Gemma 2B (it) & 0.28 & 0.31 & \cellcolor{orange!25.0} 0.50 & \cellcolor{orange!25.0} 0.37 & \cellcolor{orange!25.0} 0.63 & \cellcolor{orange!25.0} 0.63 & 0.73 & \cellcolor{orange!25.0} 0.37 \\
+Gemma 2B & \cellcolor{orange!25.0} 0.37 & \cellcolor{orange!25.0} 0.37 & \cellcolor{orange!25.0} 0.50 & 0.63 & \cellcolor{cyan!25.0} 0.14 & 0.25 & \cellcolor{orange!25.0} 0.62 & 0.45 \\
+LR & 0.03 & 0.18 & 0.79 & 0.74 & 0.03 & 0.18 & 0.79 & 0.74 \\
+GBM & 0.01 & 0.13 & 0.89 & 0.81 & 0.01 & 0.13 & 0.89 & 0.81 \\
+XGBoost & 0.00 & 0.13 & 0.90 & 0.82 & 0.00 & 0.13 & 0.90 & 0.82 \\
+\bottomrule
+\end{tabular}
+
diff --git a/results/tables/ACSMobility-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=False.tex b/results/tables/ACSMobility-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=False.tex
@@ -0,0 +1,28 @@
+\begin{tabular}{lllllllll}
+\toprule
+ & ece (mult. choice) & brier score loss (mult. choice) & roc auc (mult. choice) & accuracy (mult. choice) & ece & brier score loss & roc auc & accuracy \\
+Model &  &  &  &  &  &  &  &  \\
+\midrule
+GPT 4o mini (it) & 0.26 & 0.26 & \cellcolor{cyan!0.6} 0.57 & \cellcolor{cyan!25.0} 0.73 & 0.22 & 0.25 & 0.49 & \cellcolor{cyan!25.0} 0.73 \\
+Llama 3 70B (it) & 0.20 & \cellcolor{cyan!1.5} 0.25 & 0.57 & 0.58 & 0.05 & \cellcolor{cyan!10.7} 0.20 & 0.52 & \cellcolor{cyan!25.0} 0.73 \\
+Llama 3 70B & 0.22 & \cellcolor{cyan!4.8} 0.24 & 0.55 & 0.53 & 0.06 & \cellcolor{cyan!7.1} 0.20 & 0.53 & \cellcolor{cyan!20.7} 0.73 \\
+Llama 3 8B (it) & 0.15 & \cellcolor{cyan!13.7} 0.22 & 0.56 & \cellcolor{cyan!5.3} 0.70 & 0.11 & 0.21 & 0.49 & \cellcolor{cyan!23.4} 0.73 \\
+Llama 3 8B & \cellcolor{cyan!12.5} 0.10 & \cellcolor{cyan!21.7} 0.20 & 0.55 & \cellcolor{cyan!24.5} 0.73 & 0.14 & 0.21 & 0.51 & \cellcolor{cyan!19.1} 0.72 \\
+Mixtral 8x22B (it) & 0.40 & 0.40 & \cellcolor{orange!12.8} 0.51 & 0.39 & 0.05 & \cellcolor{cyan!17.9} 0.20 & \cellcolor{cyan!25.0} 0.54 & \cellcolor{cyan!25.0} 0.73 \\
+Mixtral 8x22B & \cellcolor{cyan!10.2} 0.11 & \cellcolor{cyan!21.2} 0.21 & 0.55 & \cellcolor{cyan!25.0} 0.73 & 0.13 & 0.22 & 0.49 & \cellcolor{cyan!25.0} 0.73 \\
+Mixtral 8x7B (it) & 0.26 & 0.26 & \cellcolor{cyan!12.8} 0.58 & \cellcolor{cyan!25.0} 0.73 & 0.11 & 0.21 & 0.51 & \cellcolor{cyan!25.0} 0.73 \\
+Mixtral 8x7B & \cellcolor{cyan!0.8} 0.14 & \cellcolor{cyan!18.9} 0.21 & 0.57 & \cellcolor{cyan!25.0} 0.73 & 0.24 & 0.25 & \cellcolor{orange!25.0} 0.48 & \cellcolor{cyan!25.0} 0.73 \\
+Mistral 7B (it) & 0.26 & 0.26 & 0.57 & \cellcolor{cyan!25.0} 0.73 & 0.17 & 0.23 & 0.49 & \cellcolor{cyan!24.5} 0.73 \\
+Mistral 7B & 0.20 & \cellcolor{cyan!8.6} 0.23 & 0.53 & \cellcolor{cyan!23.4} 0.73 & \cellcolor{orange!25.0} 0.27 & \cellcolor{orange!25.0} 0.27 & 0.50 & \cellcolor{cyan!25.0} 0.73 \\
+Yi 34B (it) & \cellcolor{cyan!18.2} 0.09 & \cellcolor{cyan!23.1} 0.20 & 0.56 & \cellcolor{cyan!15.4} 0.72 & 0.23 & 0.25 & 0.50 & \cellcolor{orange!25.0} 0.27 \\
+Yi 34B & \cellcolor{cyan!25.0} 0.07 & \cellcolor{cyan!25.0} 0.20 & 0.57 & \cellcolor{cyan!20.7} 0.73 & 0.15 & 0.23 & 0.52 & 0.44 \\
+Gemma 7B (it) & 0.25 & 0.26 & \cellcolor{cyan!25.0} 0.58 & \cellcolor{cyan!21.8} 0.73 & \cellcolor{orange!8.9} 0.25 & 0.26 & \cellcolor{orange!3.1} 0.49 & \cellcolor{cyan!22.3} 0.73 \\
+Gemma 7B & 0.41 & 0.37 & \cellcolor{orange!25.0} 0.50 & \cellcolor{orange!25.0} 0.27 & 0.19 & 0.24 & 0.49 & \cellcolor{cyan!25.0} 0.73 \\
+Gemma 2B (it) & \cellcolor{orange!25.0} 0.73 & \cellcolor{orange!25.0} 0.73 & 0.52 & \cellcolor{orange!25.0} 0.27 & \cellcolor{cyan!25.0} 0.02 & \cellcolor{cyan!25.0} 0.20 & 0.50 & \cellcolor{cyan!25.0} 0.73 \\
+Gemma 2B & 0.25 & 0.26 & 0.51 & 0.34 & \cellcolor{orange!25.0} 0.27 & \cellcolor{orange!25.0} 0.27 & 0.50 & \cellcolor{cyan!25.0} 0.73 \\
+LR & 0.02 & 0.19 & 0.61 & 0.74 & 0.02 & 0.19 & 0.61 & 0.74 \\
+GBM & 0.01 & 0.17 & 0.74 & 0.76 & 0.01 & 0.17 & 0.74 & 0.76 \\
+XGBoost & 0.00 & 0.16 & 0.74 & 0.76 & 0.00 & 0.16 & 0.74 & 0.76 \\
+\bottomrule
+\end{tabular}
+
diff --git a/results/tables/ACSMobility-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=True.tex b/results/tables/ACSMobility-table.COMBINE_PROMPTING_RESULTS=True.NUMERIC_PROMPTING=True.tex
@@ -0,0 +1,28 @@
+\begin{tabular}{lllllllll}
+\toprule
+ & ece (num) & brier score loss (num) & roc auc (num) & accuracy (num) & ece & brier score loss & roc auc & accuracy \\
+Model &  &  &  &  &  &  &  &  \\
+\midrule
+GPT 4o mini (it) & 0.22 & 0.25 & 0.49 & \cellcolor{cyan!25.0} 0.73 & 0.26 & 0.26 & \cellcolor{cyan!0.6} 0.57 & \cellcolor{cyan!25.0} 0.73 \\
+Llama 3 70B (it) & 0.05 & \cellcolor{cyan!10.7} 0.20 & 0.52 & \cellcolor{cyan!25.0} 0.73 & 0.20 & \cellcolor{cyan!1.5} 0.25 & 0.57 & 0.58 \\
+Llama 3 70B & 0.06 & \cellcolor{cyan!7.1} 0.20 & 0.53 & \cellcolor{cyan!20.7} 0.73 & 0.22 & \cellcolor{cyan!4.8} 0.24 & 0.55 & 0.53 \\
+Llama 3 8B (it) & 0.11 & 0.21 & 0.49 & \cellcolor{cyan!23.4} 0.73 & 0.15 & \cellcolor{cyan!13.7} 0.22 & 0.56 & \cellcolor{cyan!5.3} 0.70 \\
+Llama 3 8B & 0.14 & 0.21 & 0.51 & \cellcolor{cyan!19.1} 0.72 & \cellcolor{cyan!12.5} 0.10 & \cellcolor{cyan!21.7} 0.20 & 0.55 & \cellcolor{cyan!24.5} 0.73 \\
+Mixtral 8x22B (it) & 0.05 & \cellcolor{cyan!17.9} 0.20 & \cellcolor{cyan!25.0} 0.54 & \cellcolor{cyan!25.0} 0.73 & 0.40 & 0.40 & \cellcolor{orange!12.8} 0.51 & 0.39 \\
+Mixtral 8x22B & 0.13 & 0.22 & 0.49 & \cellcolor{cyan!25.0} 0.73 & \cellcolor{cyan!10.2} 0.11 & \cellcolor{cyan!21.2} 0.21 & 0.55 & \cellcolor{cyan!25.0} 0.73 \\
+Mixtral 8x7B (it) & 0.11 & 0.21 & 0.51 & \cellcolor{cyan!25.0} 0.73 & 0.26 & 0.26 & \cellcolor{cyan!12.8} 0.58 & \cellcolor{cyan!25.0} 0.73 \\
+Mixtral 8x7B & 0.24 & 0.25 & \cellcolor{orange!25.0} 0.48 & \cellcolor{cyan!25.0} 0.73 & \cellcolor{cyan!0.8} 0.14 & \cellcolor{cyan!18.9} 0.21 & 0.57 & \cellcolor{cyan!25.0} 0.73 \\
+Mistral 7B (it) & 0.17 & 0.23 & 0.49 & \cellcolor{cyan!24.5} 0.73 & 0.26 & 0.26 & 0.57 & \cellcolor{cyan!25.0} 0.73 \\
+Mistral 7B & \cellcolor{orange!25.0} 0.27 & \cellcolor{orange!25.0} 0.27 & 0.50 & \cellcolor{cyan!25.0} 0.73 & 0.20 & \cellcolor{cyan!8.6} 0.23 & 0.53 & \cellcolor{cyan!23.4} 0.73 \\
+Yi 34B (it) & 0.23 & 0.25 & 0.50 & \cellcolor{orange!25.0} 0.27 & \cellcolor{cyan!18.2} 0.09 & \cellcolor{cyan!23.1} 0.20 & 0.56 & \cellcolor{cyan!15.4} 0.72 \\
+Yi 34B & 0.15 & 0.23 & 0.52 & 0.44 & \cellcolor{cyan!25.0} 0.07 & \cellcolor{cyan!25.0} 0.20 & 0.57 & \cellcolor{cyan!20.7} 0.73 \\
+Gemma 7B (it) & \cellcolor{orange!8.9} 0.25 & 0.26 & \cellcolor{orange!3.1} 0.49 & \cellcolor{cyan!22.3} 0.73 & 0.25 & 0.26 & \cellcolor{cyan!25.0} 0.58 & \cellcolor{cyan!21.8} 0.73 \\
+Gemma 7B & 0.19 & 0.24 & 0.49 & \cellcolor{cyan!25.0} 0.73 & 0.41 & 0.37 & \cellcolor{orange!25.0} 0.50 & \cellcolor{orange!25.0} 0.27 \\
+Gemma 2B (it) & \cellcolor{cyan!25.0} 0.02 & \cellcolor{cyan!25.0} 0.20 & 0.50 & \cellcolor{cyan!25.0} 0.73 & \cellcolor{orange!25.0} 0.73 & \cellcolor{orange!25.0} 0.73 & 0.52 & \cellcolor{orange!25.0} 0.27 \\
+Gemma 2B & \cellcolor{orange!25.0} 0.27 & \cellcolor{orange!25.0} 0.27 & 0.50 & \cellcolor{cyan!25.0} 0.73 & 0.25 & 0.26 & 0.51 & 0.34 \\
+LR & 0.02 & 0.19 & 0.61 & 0.74 & 0.02 & 0.19 & 0.61 & 0.74 \\
+GBM & 0.01 & 0.17 & 0.74 & 0.76 & 0.01 & 0.17 & 0.74 & 0.76 \\
+XGBoost & 0.00 & 0.16 & 0.74 & 0.76 & 0.00 & 0.16 & 0.74 & 0.76 \\
+\bottomrule
+\end{tabular}
+