Prompt Engineering: Implement and test few prompt engineering research papers

Architecture

Key features

• Implementation and comparison of multiple prompting techniques — zero-shot, few-shot, CoT, meta, emotion prompting, and more
• LlamaIndex query engine powered by GPT-4 for prompt evaluation against a knowledge base
• Local vector store index built from Wikipedia for RAG-based prompting experiments
• text-embedding-ada-002 embeddings for semantic retrieval
• Hugging Face NER model for PII masking in prompts
• Research-backed techniques — self-consistency, emotion prompting, chain-of-thought, meta prompting, and few-shot prompting

Few code snippet from the project