
Applying Jacobian and Hessian regularization to the intermediate representations of pre-trained language models can significantly improve their generalization capabilities and uncertainty quantification.


coremsg

enhancing-generalization-and-uncertainty-quantification-in-pre-trained-language-models-through-jacobian-and-hessian-regularization


Enhancing Generalization and Uncertainty Quantification in Pre-trained Language Models through Jacobian and Hessian Regularization
