A new preprint addresses how text-to-image generators can produce varying results when the wording of a prompt is only slightly changed. The authors present a closed-loop technique that adjusts the prompt during inference, with the goal of making generated images better match the intended meaning. The revised v2 of the arXiv paper describes the approach and its evaluation.
A new arXiv paper argues that commercial text-to-image services quietly rewrite user prompts before image generation, a step people usually cannot see or turn off. Prior cultural-bias audits, the authors note, only look at finished images and treat the whole process as one pipeline, which hides where the bias enters. The work calls for examining this hidden revision stage separately.
A new arXiv paper introduces AcFlow, a technique that controls text-to-image diffusion transformers by learning conditional activation flow at inference time. The approach aims to give users adjustable control over style intensity and to suppress unwanted concepts that direct prompting handles poorly. The work targets the limits of prompt-based interfaces for generative image models.