Hallucinations
Abstract
Hallucinations are the behavior of LLMs to generate completely made-up answers that sound like normal sentences but are factually incorrect. Hallucinations make it hard to trust LLMs and come in the way of building useful applications with LLMs. In a corporate setting, it creates a business risk because the LLM can make promises about a company policy when none exists for a customer query. An LLM in education can create false narratives and mislead the learner.
Fields of Study & Subjects
Topic Modeling
Adversarial Robustness in Machine Learning
Natural Language Processing Techniques
Publication details
- Journal / source
- River Publishers eBooks
- Pages
- 41–42
- Publication date
- 12 January 2026
- Language
- EN
- Document type
- Book chapter