{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/9836bdc0398b4178bc7ae91f0ab244a9\" frameborder=\"0\" width=\"1456\" height=\"1092\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1092,"width":1456,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1092,"thumbnail_width":1456,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/9836bdc0398b4178bc7ae91f0ab244a9-00001.gif","duration":155.182,"title":"Understanding Feature Attribution in the Attention Circuit Explorer","description":"In this video, I demonstrate how to utilize recursive direct feature attribution in the Attention Circuit Explorer. By analyzing specific features in different layers, we uncover insights into how the model processes information. No specific action is requested from viewers."}