Jul 2026
What does a Bayes-filtered transformer believe? A predictive Monte Carlo approach
P predictive Monte Carlo (PMC) is used as a general interpretability tool for any BFT: using only next-token generation, PMC returns an approximation to the implicit prior and posterior over the latent task, answering the interpretive question directly in latent space.
Afiq Abdillah Effiezal Aswadi, Haotong Ma, Susan Wei
· arXiv.org · 2 citations