RAIGen: Rare Attribute Identification in Text-to-Image Generative Models

TL;DR

RAIGen is a novel unsupervised framework that identifies rare and underrepresented attributes in text-to-image diffusion models, enabling better bias detection and attribute amplification.

Contribution

It introduces RAIGen, the first method for unsupervised rare attribute discovery in diffusion models, using autoencoders and a new minority metric.

Findings

RAIGen discovers attributes beyond predefined fairness categories.

It scales to larger models like SDXL.

Supports systematic bias auditing and attribute amplification.

Abstract

Text-to-image diffusion models achieve impressive generation quality but inherit and amplify training-data biases, skewing coverage of semantic attributes. Prior work addresses this in two ways. Closed-set approaches mitigate biases in predefined fairness categories (e.g., gender, race), assuming socially salient minority attributes are known a priori. Open-set approaches frame the task as bias identification, highlighting majority attributes that dominate outputs. Both overlook a complementary task: uncovering rare or minority features underrepresented in the data distribution (social, cultural, or stylistic) yet still encoded in model representations. We introduce RAIGen, the first framework, to our knowledge, for un-supervised rare-attribute discovery in diffusion models. RAIGen leverages Matryoshka Sparse Autoencoders and a novel minority metric combining neuron activation frequency with semantic distinctiveness to identify interpretable neurons whose top-activating images reveal underrepresented attributes. Experiments show RAIGen discovers attributes beyond fixed fairness categories in Stable Diffusion, scales to larger models such as SDXL, supports systematic auditing across architectures, and enables targeted amplification of rare attributes during generation.