Improving Multilingual Language Models by Aligning Representations through Steering

Open in new window