AI.info

Fuyuko Kido

Explore Fuyuko Kido on AI.info.

• OpenLVLM-MIA: A Controlled Benchmark Revealing the Limits of Membership Inference Attacks on Large Vision-Language Models