Data: Fine-tuning and RAG
Abstract
LLM carries its pre-trained data knowledge in its parameters. GenAI showed us the promise of LLM to create context-aware conversational chatbots to engage with customers. As enterprises began exploring this capability, companies asked for the chatbot to utilize their data sources. For example, a customer service chatbot can serve customers only when it has learned from past customer chat archives to understand top issues and solutions. This brought the innovation of RAG (retrieval-augmented generation) to allow pre-training with non-parametric memory ( Lewis et al., 2020 ).
Fields of Study & Subjects
Topic Modeling
Information Retrieval and Search Behavior
Multimodal Machine Learning Applications
Publication details
- Journal / source
- River Publishers eBooks
- Pages
- 43–44
- Publication date
- 12 January 2026
- Language
- EN
- Document type
- Book chapter