1 paper
Hang Zhang, Warren J. Gross
Not all training samples contribute equally to large language model fine-tuning. Selecting informative training samples can reduce the computational cost while preserving downstrea…