In this notebook, we'll be implementing the papers written on query translation to improve RAG performance. Traditional RAG approaches may not always yield good results and may hallucinate even when the query given to them is ambiguous or entirely out of scope. Query translation helps us in this sense by reducing hallucinations and yielding improved results.
- Least-to-Most Prompting Enables Complex Reasoning in Large Language Models, https://arxiv.org/pdf/2205.10625.pdf
- Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions, https://arxiv.org/pdf/2212.10509.pdf
https://www.kaggle.com/code/deeepsig/implementation-of-query-translation-papers/notebook