Reducing Hallucinations in Vision-Language Models via Latent Space Steering

Open in new window