Modern Large Language Models such as GPT-4, LLaMA, and DeepSeek utilize transformer-based neural architectures trained to estimate conditional probabilities of token sequences.

Authors

• Person: Not available Organization: Frontiers
  Survey and analysis of hallucinations in large language models

Sources

• Survey and analysis of hallucinations in large language models www.frontiersin.org Frontiers via serper

Referenced by nodes (1)

• Large Language Models concept