Large Language Models

LLMs are a primary target for AI Red Teaming. Understanding their architecture (often Transformer-based), training processes (pre-training, fine-tuning), and capabilities (text generation, summarization, Q&A) is essential for identifying vulnerabilities like prompt injection, jailbreaking, data regurgitation, and emergent harmful behaviors specific to these large-scale models.

Learn more from the following resources:

• @article@What is an LLM (large language model)?
• @guide@ChatGPT For Everyone
• @guide@What Are Large Language Models? A Beginner's Guide for 2025