ResearchBenchmarksVisual Evidence Quality Evaluation on LLaVA-W In-the-Wild (subset of 500 samples)Follow81.2AIM-CoT Win RateAIM-CoT77.1479.1781.283.23Sep 30, 2025Evaluation ResultsMethodMethodLinksAIM-CoT Win RateTie RateICoT Win RateAIM-CoTSelection Strategy=Inf...Selection Strategy=Info. Gain-driven AVP, Judge=GPT-4v2025.0981.213.65.2ICoTSelection Strategy=Att...Selection Strategy=Attention-driven Top-K, Judge=GPT-4v2025.0981.213.65.2