Skip to yearly menu bar Skip to main content


LANCE: Stress-testing Visual Models by Generating Language-guided Counterfactual Images

Viraj Prabhu · Sriram Yenamandra · Sriram Yenamandra · Prithvijit Chattopadhyay · Judy Hoffman

Great Hall & Hall B1+B2 (level 1) #1606


We propose an automated algorithm to stress-test a trained visual model by generating language-guided counterfactual test images (LANCE). Our method leverages recent progress in large language modeling and text-based image editing to augment an IID test set with a suite of diverse, realistic, and challenging test images without altering model weights. We benchmark the performance of a diverse set of pre-trained models on our generated data and observe significant and consistent performance drops. We further analyze model sensitivity across different types of edits, and demonstrate its applicability at surfacing previously unknown class-level model biases in ImageNet. Code is available at

Chat is not available.