LLMs- Consider Hallucinatory Unless Proven Otherwise

Explore the path towards a hallucination-free future in this 22-minute conference talk by Shayak Sen, CTO of TruEra. Delve into the world of Large Language Models (LLMs) and their generalization capabilities. Discover the RAG Triad and its significance in improving LLM performance. Learn about TruLens and the importance of feedback functions in enhancing model accuracy. Gain insights into query planning strategies and best practices for implementing these techniques. The talk concludes with a Q&A session, providing additional insights into the topic. This presentation offers valuable knowledge for those interested in advancing LLM technology and reducing hallucinations in AI-generated content.

Introduction
LLMs and Generalization
The RAG Triad
TruLens
Feedback Functions
Getting This Right
Query Planning
Conclusion
Q&A