Skip to main content

Large Language Model (LLM)

A type of AI system trained on vast amounts of text to understand and generate human language.

Definition​

A Large Language Model is an AI program that has "read" enormous amounts of text—books, articles, websites, code—and learned patterns in how language works. When you ask it a question or give it a prompt, it predicts what text should come next based on those patterns.

Why it matters for humanities​

LLMs are the technology behind ChatGPT, Claude, and similar tools. For humanities researchers, they can:

  • Summarize long documents or archives
  • Translate between languages (including historical forms)
  • Generate drafts, outlines, or brainstorming material
  • Analyze patterns in text corpora
  • Explain technical concepts in accessible language

Key characteristics​

AspectWhat it means
"Large"Billions of parameters (internal settings)
"Language"Trained primarily on text data
"Model"A statistical representation, not true understanding

Common LLMs​

  • GPT-4 (OpenAI) — Powers ChatGPT
  • Claude (Anthropic) — Known for longer documents and nuance
  • Gemini (Google) — Integrated with Google services
  • Llama (Meta) — Open-source, can run locally

Limitations​

LLMs don't "know" things the way humans do. They:

  • Can produce confident-sounding errors ("hallucinations")
  • Have knowledge cutoffs (don't know recent events)
  • May reflect biases present in training data
  • Cannot verify factual claims or cite sources reliably
  • Token
  • Context window
  • Hallucination