Davide Baldelli

This work was carried out as part of MATS 10.0 under the supervision of Mirko Bronzi and Damiano Fornasiere. The full paper can be found here. TL;DR * Probes on activations are widely used to detect misbehavior in language models, but they rely on the assumption that models cannot control...

Pseudo-random generation solved the problem of sampling from mathematical distributions on computers. Faced with a similar need in natural language, it feels natural to simply ask an LLM: "Name a random city" but this leads to the disappointing discovery that language models are heavily biased and suffer from mode collapse....