 initial model & dataset & Llama-2-70b-chat-hf & Mixtral-8x7B-Instruct-v0.1 & Qwen1.5-72B-Chat \\
 & \scalebox{0.9}[1]{MathGen} & \cellcolor[RGB]{240,230,140}{67.5} & \cellcolor[RGB]{240,230,140}{72.4} & \cellcolor[RGB]{240,230,140}{76.3} & \\
 & \scalebox{0.9}[1]{FgFactV} & \cellcolor[RGB]{240,230,140}{55.5} & \cellcolor[RGB]{240,230,140}{52.1} & \cellcolor[RGB]{240,230,140}{64.3} & \\
 & \scalebox{0.9}[1]{AnsCls} & \cellcolor[RGB]{240,230,140}{59.4} & \cellcolor[RGB]{240,230,140}{60.4} & \cellcolor[RGB]{240,230,140}{55.7} & \\
 & \scalebox{0.9}[1]{MathGen} & \cellcolor[RGB]{240,230,140}{84.4} & \cellcolor[RGB]{240,230,140}{87.0} & \cellcolor[RGB]{240,230,140}{93.5} & \\
 & \scalebox{0.9}[1]{FgFactV} & \cellcolor[RGB]{240,230,140}{80.5} & \cellcolor[RGB]{240,230,140}{96.3} & \cellcolor[RGB]{240,230,140}{73.6} & \\
 & \scalebox{0.9}[1]{AnsCls} & \cellcolor[RGB]{240,230,140}{73.6} & \cellcolor[RGB]{240,230,140}{83.6} & \cellcolor[RGB]{240,230,140}{70.0} & \\
