Główne pojęcia
Efficiently leverage LLMs for cost-effective data annotation in NLP.
Statystyki
"Our results indicate that the LlamBERT approach slightly compromises on accuracy while offering much greater cost-effectiveness."
"The large version of RoBERTa performed the best on all 4 training scenarios, reaching a state-of-the-art accuracy of 96.68%."
"We can estimate that labeling the entirety of IMDb’s 7.816 million movie reviews would take about 48h 28m with roberta-large."
Cytaty
"We believe that such resource-efficient solutions can foster sustainable development and environmental stewardship."
"Combining the LlamBERT technique with fine-tuning on gold-standard data yielded the best results in both cases."
"Our code is available on GitHub."