Turning an LLM into a classifier

Asking an LLM "is this review positive" works, and works immediately. When you should move to a small model is a measurable question.

Sources

• Brown, T. et al. 2020 · Language Models are Few-Shot Learners · NeurIPS 2020
• Hsieh, C.-Y. et al. 2023 · Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes · ACL 2023 Findings
• Zhao, Z. et al. 2021 · Calibrate Before Use: Improving Few-Shot Performance of Language Models · ICML 2021
• Bucila, C., Caruana, R. & Niculescu-Mizil, A. 2006 · Model Compression · KDD 2006