arXivRuihan Li, Jiyang Tan, Kailin Jiang, Huining Li, Hengyang Lu, Yu Huang, Qian Li, Yuntao DuTue, Aug 4, 2026, 8:03 AM PDT
score 17.0
New benchmark tests when AI image models make things up
Original: KnowHal: A Knowledge-Driven Benchmark for Comprehensive Multimodal Hallucination Evaluation
Source: arxiv.org ↗
Writing ELI5 summary…