Yossi Gandelsman

发表论文 38 篇 · 总被引 2282 次 · h-index 19

代表论文

• Interpreting CLIP's Image Representation via Text-Based Decomposition (2023 · International Conference on Learning Representations · 被引 212)
• Interpreting and Editing Vision-Language Representations to Mitigate Hallucinations (2024 · International Conference on Learning Representations · 被引 105)
• Vision Transformers Don't Need Trained Registers (2025 · arXiv.org · 被引 41)
• Interpreting the Second-Order Effects of Neurons in CLIP (2024 · International Conference on Learning Representations · 被引 39)
• Interpreting the Weight Space of Customized Diffusion Models (2024 · Neural Information Processing Systems · 被引 38)
• Steering CLIP's vision transformer with sparse autoencoders (2025 · arXiv.org · 被引 33)