paper
StylisticBias: A Few Human Visual Cues Drive Most Social Biases in MLLMs
paperactiveprovisional
stylisticbias-a-few-human-visual-cues-drive-most-social-biases-in-mllms-a9c822ef·1 events·first seen 47h agoAliases: StylisticBias: A Few Human Visual Cues Drive Most Social Biases in MLLMs
Co-occurring entities
More like this (12)
StylisticBiasgender bias in VLMsBeyond Third-Person Audits: Situated Interaction Auditing for User-Centered LLM Bias ResearchFrom Self to Other: Evaluating Demographic Perspective-Taking in LLM Hate Speech AnnotationWatch, Remember, Reason: Human-View Video Understanding with MLLMsContagion Networks: Evaluator Bias Propagation in Multi-Agent LLM SystemsThe Masked Advantage: Uncovering Local-Language Access to Cultural Knowledge in LLMsPerceptual Judgment BiasSocial Gaze ConsistencyModeling Complex Behaviors: Multi-Personality Composition and Dynamic Switching in Vision-Language ModelsWhere Does the Answer Come From? Benchmarking View-Level Visual Evidence Identification in Multi-View MLLMs for Autonomous DrivingGaze Heads: How VLMs Look at What They Describe
Recent events (1)
StylisticBias benchmark reveals a small set of visual cues drives most social bias in MLLMs
Researchers introduce StylisticBias, a controlled benchmark of ~25K photorealistic face images with single-attribute variations designed to isolate how specific visual cues shift social judgments in multimodal LLMs. Evaluating six MLLMs across 25 binary social judgment scenarios, they find that age and body type dominate identity-level effects, while fashion style drives the largest attribute-level shifts, with ~15 attributes accounting for ~80% of total bias variation. The benchmark is released publicly on GitHub and Hugging Face, enabling fine-grained bias auditing of multimodal models.