Hallucination

How it is used

Teams measure, detect and reduce hallucinations, often with retrieval (RAG), citations and automatic checks. The verb is hallucinate: 'the model hallucinated a paper that doesn't exist'. Some researchers prefer the word 'confabulation'.

Example sentences

1. This answer is a hallucination: the two court cases it cites never existed.

2. Grounding answers in our own documents cut hallucinations by more than half.

Common collocations

• reduce hallucinations
• hallucination rate
• detect hallucinations
• the model hallucinated

Related terms

• LLMShort for large language model: a very large neural network trained on huge amounts of text to predict the next token, which lets it answer questions, write, summarize and code.
• RAGShort for retrieval-augmented generation: a method in which a system first searches a document collection for relevant passages and then gives them to an LLM to write the answer.
• guardrailA rule or check placed around a model to stop harmful, off-topic or unsafe inputs and outputs, for example a filter that blocks personal data in answers.
• alignmentMaking an AI system's behavior match human intentions and values, so that it is helpful and honest and avoids causing harm. Also the research field that studies this.

This word in other fields

The same English word can have a different meaning in another field:

• hallucination · Artificial Intelligence & Machine Learning

More terms: Machine Learning and Data Processing

• artificial intelligence
• machine learning
• model
• dataset
• training data
• label
• train
• supervised learning
• neural network
• deep learning
• loss function
• overfitting