← Search

Lennie Wells

2 accepted papers

2026

Same Question, Different Lies: Cross-Context Consistency (C³) for Black-Box Sandbagging Detection

ICML 2026poster

As language models grow more capable, accurate capability evaluation becomes essential for safety decisions. If models can deliberately underperform on dangerous capability evaluations---a behavior known as \emph{sandbagging}---they may evade safety measures designed for their true capability level.…

Cited by 0SourceScholar
2024

Unconstrained Stochastic CCA: Unifying Multiview and Self-Supervised Learning

ICLR 2024poster

The Canonical Correlation Analysis (CCA) family of methods is foundational in multiview learning. Regularised linear CCA methods can be seen to generalise Partial Least Squares (PLS) and be unified with a Generalized Eigenvalue Problem (GEP) framework. However, classical algorithms for these linear…

Cited by 2SourcePDFScholar