Notes of ChatGPT Prompt Engineering for Developers
For ease of referring and checking in the future, I summarized this outline for the short course ChatGPT Prompt Engineering for Developers at DeepLearning.AI. Introduction Introducing the general background of natural language process (NLP) with deep learning (DL). Large language models (LLMs) [I think here the term refers to GPT-like LLMs] were categorized into two types. Base LLM Predicts the next word based on the training data Instruction Tuned LLM Tries to follow instructions (Fine-tuning based on instructions and making good attempts at following those instructions) RLHF: Reinforcement Learning with Human Feedback (arXiv paper) Helpful, Honest, Harmless [at least we hope so] ...