model,baseline,n,answer_in_survivor_set,accuracy,accuracy_within_survivors Qwen_Qwen2.5-VL-3B-Instruct,0.00,764,0.6898,0.3626,0.5256 Qwen_Qwen2.5-VL-3B-Instruct,0.05,764,0.6963,0.3743,0.5376 Qwen_Qwen2.5-VL-3B-Instruct,0.10,764,0.7042,0.3783,0.5372 Qwen_Qwen2.5-VL-3B-Instruct,0.15,764,0.7134,0.3796,0.5321 Qwen_Qwen2.5-VL-3B-Instruct,0.20,764,0.7186,0.3809,0.5301 Qwen_Qwen2.5-VL-3B-Instruct,0.25,764,0.7094,0.3704,0.5221 Qwen_Qwen2.5-VL-7B-Instruct,0.00,764,0.6898,0.4031,0.5844 Qwen_Qwen2.5-VL-7B-Instruct,0.05,764,0.7003,0.394,0.5626 Qwen_Qwen2.5-VL-7B-Instruct,0.10,764,0.699,0.3979,0.5693 Qwen_Qwen2.5-VL-7B-Instruct,0.15,764,0.6937,0.4018,0.5792 Qwen_Qwen2.5-VL-7B-Instruct,0.20,764,0.7016,0.377,0.5373 Qwen_Qwen2.5-VL-7B-Instruct,0.25,764,0.7068,0.377,0.5333 Qwen_Qwen3.5-4B,0.00,764,0.623,0.3914,0.6282 Qwen_Qwen3.5-4B,0.05,764,0.6178,0.3901,0.6314 Qwen_Qwen3.5-4B,0.10,764,0.6113,0.3822,0.6253 Qwen_Qwen3.5-4B,0.15,764,0.6191,0.3874,0.6258 Qwen_Qwen3.5-4B,0.20,764,0.6099,0.3796,0.6223 Qwen_Qwen3.5-4B,0.25,764,0.6034,0.3848,0.6377 Qwen_Qwen3.5-4B-Base,0.00,764,0.6178,0.4424,0.7161 Qwen_Qwen3.5-4B-Base,0.05,764,0.606,0.4529,0.7473 Qwen_Qwen3.5-4B-Base,0.10,764,0.6073,0.4503,0.7414 Qwen_Qwen3.5-4B-Base,0.15,764,0.6126,0.4542,0.7415 Qwen_Qwen3.5-4B-Base,0.20,764,0.6178,0.4503,0.7288 Qwen_Qwen3.5-4B-Base,0.25,764,0.6204,0.4398,0.7089