arXivParamananda Bhaskar, Naquee Rizwan, Daksh Jogchand, Saurabh Kumar Pandey, Animesh MukherjeeFri, May 29, 2026, 7:27 AM PDT

score 14.6

New benchmark exposes vision-language model failures on hateful memes

Original: FBHM: Functional Benchmarking and Steering of VLMs for Hateful Meme Detection

Writing ELI5 summary…