3 ms·Medical large language models are vulnerable to data-poisoning attacks2 points by gnabgib 2y ago