2026-04-29 00:00 UTCOriginal source2 min readUpdated: 2026-06-27 00:25 UTC

DSO: Direct Steering Optimization for Bias Mitigation

Apple ML Research introduces DSO (Direct Steering Optimization), a reinforcement learning-based method that learns linear transformations to steer model activations, effectively mitigating bias in VLMs and LLMs. It achieves state-of-the-art fairness-capabilities trade-off with inference-time control.

SourceApple Machine Learning Research

DSO: Direct Steering Optimization for Bias Mitigation - Apple Machine Learning Research

Machine Learning Research

Open MenuClose Menu

Overview

Research Highlights

Publications

Events

Work with us

research area Fairness, research area Methods and Algorithmsconference CVPR

content type paperpublished April 2026

DSO: Direct Steering Optimization for Bias Mitigation

AuthorsLucas Monteiro Paes‡, Nivedha Sivakumar‡, Oliver Wang†‡**, Masha Fedzechkina, Barry-John Theobald, Luca Zappella, Nicholas Apostoloff

View publication

View source code (GitHub)

Copy Bibtex

Generative models are often deployed to make decisions on behalf of users, such as vision-language models (VLMs) identifying which person in a room is a doctor to help visually impaired individuals. Yet, VLM decisions are influenced by the perceived demographic attributes of people in the input, which can lead to biased outcomes like failing to identify women as doctors. Moreover, when reducing bias leads to performance loss, users may have varying needs for balancing bias mitigation with overall model capabilities, highlighting the demand for methods that enable controllable bias reduction during inference. Activation steering is a popular approach for inference-time controllability that has shown potential in inducing safer behavior in large language models (LLMs). However, we observe that current steering methods struggle to correct biases, where equiprobable outcomes across demographic groups are required. To address this, we propose Direct Steering Optimization (DSO) which uses reinforcement learning to find linear transformations for steering activations, tailored to mitigate bias while maintaining control over model performance. We demonstrate that DSO achieves state-of-the-art trade-off between fairness and capabilities on both VLMs and LLMs, while offering practitioners inference-time control over the trade-off. Overall, our work highlights the benefit of designing steering strategies that are directly optimized to control model behavior, providing more effective bias intervention than methods that rely on pre-defined heuristics for controllability.

† Carnegie Mellon University

‡ Equal contribution

** Work done while at Apple