AI.info

Hayato Yamana

Explore Hayato Yamana on AI.info.

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