LaMDA: Large Model Fine-Tuning via Spectrally Decomposed Low-Dimensional Adaptation.

Seyedarmin Azizi, Souvik Kundu, Massoud Pedram

Browse the full EMNLP paper archive.