Beyond Guardrails: Defending LLMs Against Sophisticated Attacks
The Data Exchange with Ben Lorica
00:00
Navigating the Complexities of RAG and LLM Biases
This chapter explores the complexities of Retrieval-Augmented Generation (RAG) and the persistent biases in Large Language Models (LLMs). It also examines the vulnerabilities to prompt injections in both proprietary and open-weight models, shedding light on their differing responses and knowledge capacities.
Transcript
Play full episode
