Seeing is Not Understanding: A Benchmark on Perception-Cognition Disparities in Large Language Models
This addresses a gap in evaluating emotion understanding for researchers and developers of MLLMs, though it is incremental as it focuses on a specific domain.
The paper tackles the problem of inadequate assessment of multimodal large language models' ability to understand complex human emotions by introducing EmoBench-Reddit, a hierarchical benchmark with 350 curated samples, and finds that current models struggle with advanced cognition tasks despite excelling in perception.
With the rapid advancement of Multimodal Large Language Models (MLLMs), they have demonstrated exceptional capabilities across a variety of vision-language tasks. However, current evaluation benchmarks predominantly focus on objective visual question answering or captioning, inadequately assessing the models' ability to understand complex and subjective human emotions. To bridge this gap, we introduce EmoBench-Reddit, a novel, hierarchical benchmark for multimodal emotion understanding. The dataset comprises 350 meticulously curated samples from the social media platform Reddit, each containing an image, associated user-provided text, and an emotion category (sad, humor, sarcasm, happy) confirmed by user flairs. We designed a hierarchical task framework that progresses from basic perception to advanced cognition, with each data point featuring six multiple-choice questions and one open-ended question of increasing difficulty. Perception tasks evaluate the model's ability to identify basic visual elements (e.g., colors, objects), while cognition tasks require scene reasoning, intent understanding, and deep empathy integrating textual context. We ensured annotation quality through a combination of AI assistance (Claude 4) and manual verification.We conducted a comprehensive evaluation of nine leading MLLMs, including GPT-5, Gemini-2.5-pro, and GPT-4o, on EmoBench-Reddit.