Xujia Wang

low-rank adaptation (2) parameter-efficient fine-tuning (2) large language model (2) multi-task learning (1) knowledge transfer (1) knowledge distillation (1) cross-lingual transfer (1) model merging (1) language model (1) mixture of expert (1) supervised fine-tuning (1) parameter importance (1) high-rank adaptation (1) knowledge grounding (1) multilingual language model (1) overfitting prevention (1) knowledge fusion (1) subnetwork localization (1) gradient sparsity (1) subnetwork optimization (1)