1 paper
Eric Khiu, Hasti Toossi, David Anugraha +6
Fine-tuning and testing a multilingual large language model is expensive and challenging for low-resource languages (LRLs). While previous studies have predicted the performance of…