Latest Articles for Technology Safety

Research highlights methods to detect backdoor attacks in fine-tuning language models.

Research reveals vulnerabilities in AI image generators from prompt manipulation.

A database to combat backdoor defects in deep learning models.

Ensemble learning improves safety filters in control systems, enhancing decision-making for technology.

Granite Guardian safeguards AI conversations from harmful content effectively.

A new method ensuring language models remain safe while performing effectively.

Setting rules for AI safety while avoiding sneaky tricks.