Prompt injection evals on six small models. They acted on hidden instructions from never to half the time, and zero data left. Gate the boundary.
2 posts tagged
Prompt injection evals on six small models. They acted on hidden instructions from never to half the time, and zero data left. Gate the boundary.
Mutation testing for infrastructure code, done by hand across fifteen repositories. 153 planted bugs, and four green tests that could never have failed.