Medical large language models are vulnerable to data-poisoning attacksDaniel Alexander Alber, Eric K. Oermann, Anton Alyakin et al.|Nature Medicine|2025Cited by 150