Prototypicality Bias Reveals Blindspots in Multimodal Evaluation Metrics

TL;DR

This paper uncovers a bias in multimodal evaluation metrics where they favor prototypical images over semantically correct but non-prototypical ones, proposing a new metric to improve robustness.

Contribution

The authors introduce ProtoBias, a benchmark for evaluating prototypicality bias, and propose ProtoScore, a more robust metric that reduces failure rates and improves semantic evaluation.

Findings

Widely used metrics often misrank prototypical adversarial images.

Human judgments favor semantic correctness over prototypes.

ProtoScore significantly outperforms existing metrics in robustness.

Abstract

Automatic metrics are now central to evaluating text-to-image models, often substituting for human judgment in benchmarking and large-scale filtering. However, it remains unclear whether these metrics truly prioritize semantic correctness or instead favor visually and socially prototypical images learned from biased data distributions. We identify and study prototypicality bias as a systematic failure mode in multimodal evaluation. We introduce a controlled contrastive benchmark ProtoBias (Prototypical Bias), spanning Animals, Objects, and Demography images, where semantically correct but non-prototypical images are paired with subtly incorrect yet prototypical adversarial counterparts. This setup enables a directional evaluation of whether metrics follow textual semantics or default to prototypes. Our results show that widely used metrics, including CLIPScore, PickScore, and VQA-based scores, frequently misrank these pairs, while even LLM-as-Judge systems exhibit uneven robustness in socially grounded cases. Human evaluations consistently favour semantic correctness with larger decision margins. Motivated by these findings, we propose ProtoScore, a robust 7B-parameter metric that substantially reduces failure rates and suppresses misranking, while running at orders of magnitude faster than the inference time of GPT-5, approaching the robustness of much larger closed-source judges.