2 papers indexed here
We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.
Not the right person? Other researchers publish under this name.
A PAC-Bayes excess-risk bound is proved that, for the first time, jointly accounts for prediction-space redundancy and meta-learner capacity in heterogeneous prediction pools.
For harmful prompts, the projection of the last-token hidden state onto a learned harm direction rises monotonically with transformer depth, whereas benign prompts remain flat or oscillatory, supporting the projection sequence as a reproducible structured signal.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
