一作ACL 2026Poster· Natural Language Processing
Can Spectral-Clipping Enable Better Learning While Forgetting Less for Low-Rank Adaptation?
Hyowon Wi · Noseong Park
摘要
In recent years, low-rank adaptation (LoRA) has emerged as a significant paradigm that freezes pre-trained weights and introduces small, learnable adapters instead of fine-tuning the full set of parameters. In this work, we uncover several key insights regarding the \textit{singular} components o...