All the articles with the tag "training-data-poisoning".
How poisoned training data plants backdoors and bias in language models, where it enters the pipeline, and the defenses that catch it early.
