1 paper
Yang Song, Yixuan Zhang, Lingfa Meng +5
Parameter-efficient adaptation has made LLMs practical for domain prediction, but standard LoRA still relies on a static low-rank update and does not expose the latent interactions…