1. Scaling laws

2. Emergent abilities

3. Instruction tuning

4. Reinforcement Learning from Human Feedback (RLHF)

5. Prompt engineering

6. Chain-of-thought prompting

7. In-context learning

8. Retrieval-Augmented Generation (RAG)

9. Hallucination & factuality

10. Model interpretability & explainability

11. Parameter-efficient fine-tuning (LoRA, adapters, prefix tuning)

12. Quantization & model compression