Datasets:
| input,model,participants,accuracy,macro_f1,coverage,selective_ordinal_mae,selective_within_one,elevated_recall | |
| direct_video,Keye-VL-1.5-8B,704,0.1747,0.0744,1.0,1.1023,0.75,0.0 | |
| direct_video,Qwen3-VL-8B-Instruct,704,0.2514,0.1443,1.0,0.9702,0.8026,0.0 | |
| privte_text,Qwen3.6-Flash,704,0.3935,0.2094,0.9616,0.7208,0.8833,0.0 | |