IARPA wants new options to spot when large language models exhibit potentially harmful behavior
The intelligence community’s primary research arm is exploring new ways to detect and combat vulnerabilities, biases and threats associated with emerging generative AI and large language model…