Md. Al Amin
Titles as the sources gave them. In 1 entry, with 0 checked quotations.
In the Base
Every entry they appear in.
Clean Sheets and Checkpoints: Evidence Complexity for Certifying Sequential Agent Risk
According to the abstract, four authors model dangerous-capability tests of AI agents as a chain of milestones. They ask how many zero-success audit results can certify, at a stated confidence level, that an agent's chance of completing the chain stays below a set limit, and derive bounds.