Visual Anagrams Reveal Hidden Differences in Holistic Shape Processing Across Vision Models

Open in new window