179 questions
No questions match those filters.
← All questions
Security & SafetyMedium· Prompt Injection· Data Poisoning· Mechanistic Interpretability
Your report told the security team the generator copies...
This is one of the questions in the full AI/ML interview bank. Pro unlocks all 1789 questions; Premium includes the same bank plus the highest daily Practice limit.
See plansThe circuit predicted this behavior exactly, confirming the mechanism. The problem is that mechanism is not equivalent to warrant. To mitigate this, one must implement credibility scoring at retrieval or modify attention on low-credibility spans, rather than attempting to fine-tune the generator.