LLM on a Budget: Active Knowledge Distillation for Efficient Classification of Large Text Corpora
Explore M-RARU, an active learning framework that leverages uncertainty sampling to minimize API costs while maintaining high classification accuracy in larg...