                 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}{55.4} & \cellcolor[RGB]{240,230,140}{46.9} & \cellcolor[RGB]{240,230,140}{31.9} &                                 \\
                               &     \scalebox{0.9}[1]{FgFactV} & \cellcolor[RGB]{240,230,140}{65.4} & \cellcolor[RGB]{240,230,140}{42.9} & \cellcolor[RGB]{240,230,140}{25.1} &                                 \\
                               &      \scalebox{0.9}[1]{AnsCls} & \cellcolor[RGB]{240,230,140}{65.9} & \cellcolor[RGB]{240,230,140}{39.3} & \cellcolor[RGB]{240,230,140}{16.5} &                                 \\
                               &     \scalebox{0.9}[1]{MathGen} & \cellcolor[RGB]{240,230,140}{66.8} & \cellcolor[RGB]{240,230,140}{56.7} & \cellcolor[RGB]{240,230,140}{51.6} &                                 \\
                               &     \scalebox{0.9}[1]{FgFactV} & \cellcolor[RGB]{240,230,140}{80.6} & \cellcolor[RGB]{240,230,140}{35.5} & \cellcolor[RGB]{240,230,140}{18.4} &                                 \\
                               &      \scalebox{0.9}[1]{AnsCls} & \cellcolor[RGB]{240,230,140}{50.2} & \cellcolor[RGB]{240,230,140}{30.8} & \cellcolor[RGB]{240,230,140}{4.6} &                                 \\
