AI Safety

2 papers touch this idea.

Papers

AI models can't tell when a hacker put words in their mouth

A hidden AI bias you can't see in its text? Force it to leak.

Related concepts

Ideas that show up alongside AI Safety.

Adversarial Attacks Bias Detection In Language Models Fisher Information Knowledge Distillation LLM Introspection Model Auditing Model Interpretability Model Supply Chain Security Parameter-Efficient Finetuning Prefix Tuning Refusal Behavior Self-Report Reliability