When Steering Works: Predicting Debiasing Benefit from Denoiser Responses

Semantic steering can reduce stereotypes in text-to-image generation, but it can also change content that the prompt asks for. Can we predict which prompts will benefit before decoding their images? We introduce Directional Gain ($G$), which compares two denoiser responses at the same latent state: the change caused by steering and the change caused by replacing a stereotype prompt with its counterfactual. We measure agreement between these directions relative to random edits. Across SD3.5, FLUX, and Qwen-Image, with 1,295 test prompts per model, $G$ predicts improvement better than response size. At the top 20%, selection by $G$ gives larger stereotype reductions and usually smaller prompt-matching losses. Its advantage in combined-score improvement over response size is 4.7–20.7 points per selected prompt under Qwen3-VL; GPT-5.6 also favors $G$ in all three models. Averaged over six single-probe choices, one probe retains 89–93% of the mean improvement obtained by six-probe selection. It uses fewer denoiser evaluations than generating a clean–steered image pair and requires no image decoding or judging. With the steering settings fixed, the ranking also selects prompts that benefit on new generation seeds.