How does prompting differ from fine-tuning a language model?
Instruction: Give a clear explanation, a practical example, and the main limitation.
Context: Compare changes to inference-time inputs with changes to model weights, including the data and evaluation each needs.
Updated
Official answer available
Read the opening below, then unlock the full answer and practical guidance.
Prompting supplies instructions and examples at inference time; it does not update the model weights. Fine-tuning changes weights using training examples...