Answer 5/8 Questions of 6 marks each
(a)Define: Prompt, Prompt Engineering (b.)Enumerate the 5 principles of prompting. Give a one line brief for each principle.
(c)Enumerate 5 issues/problems you 'd want attempt to fix, if you were building a product name generator with thousands of users.
2.Define: Text generation models, Token, Tokenization. Enumerate the 2 phases of reliable LLM development. Brief them.
3.Brief about "Vector Representations: The Numerical Essence of Language".
4.Define (a)"word vectors/word embeddings" in NLP (b)Transformer/Transformer architecture (c)Brief about "Transformer Architecture: Orchestrating Contextual Relationships"
5.Brief about "Probabilistic Text Generation: The Decision Mechanism".
6.Contrast: "Historical Underpinnings: The Rise of Transformer Architectures-OpenAI’s Generative Pretrained Transformers Vs. Google’s Gemini Vs. Meta’s Llama and Open Source.".
7.(a)Enumerate 3 different benefits of .yml files over .json files. (b)Brief about "Filtering YAML Payloads". Depict Figure 3.1 for using an LLM to determine the control flow of application instead of code.
8. Brief about "Diverse Format Generation with ChatGPT".