Hallucination
AI hallucination is the phenomenon where a language model generates confident-sounding but factually incorrect, fabricated, or nonsensical information that is not grounded in the input or training data.
理解する Hallucination
Hallucination is one of the most significant challenges with deploying large language models. LLMs do not retrieve facts from a database; they predict the next token based on statistical patterns learned during training. When asked about something outside their training data or at the edges of their knowledge, models sometimes generate plausible-sounding but false information with apparent confidence. Hallucinations range from subtle factual errors (wrong dates, incorrect statistics) to complete fabrications (invented citations, non-existent people, made-up product features). The danger is that the model's confident tone gives no indication that the information is incorrect. Several factors increase hallucination risk: asking for very specific facts the model was not trained on, requesting information about real but obscure entities, asking leading questions that imply false premises, and giving the model insufficient context. Low-temperature sampling reduces but does not eliminate hallucinations. The primary mitigation strategies are grounding and RAG. Grounding means providing the model with source documents and instructing it to base its answers only on that content. RAG retrieves relevant documents before generation, giving the model accurate information to reference. These techniques are particularly effective for factual query tasks.
GAIAの活用方法 Hallucination
GAIA minimizes hallucination by grounding responses in retrieved data. When answering questions about your emails, tasks, or projects, GAIA first retrieves relevant content from ChromaDB and injects it into the prompt as context. The LLM is instructed to base its response on the retrieved data rather than general knowledge, dramatically reducing the risk of fabricated information in productivity-critical contexts.
関連概念
Retrieval-Augmented Generation (RAG)
Retrieval-Augmented Generation (RAG) is a technique that enhances LLM responses by first retrieving relevant documents or data from an external knowledge base and injecting that context into the model's prompt.
Prompt Engineering
Prompt engineering is the practice of designing and refining inputs to AI language models to reliably elicit desired outputs, shaping model behavior without modifying the underlying weights.
大規模言語モデル（LLM）
大規模言語モデル（LLM）は、膨大なテキストデータでトレーニングされた人工知能モデルであり、人間のような流暢さで言語を理解、生成、推論できます。
Context Window
The context window is the maximum number of tokens a language model can process in a single inference call, encompassing the system prompt, conversation history, retrieved documents, and generated output.
Large Language Model (LLM)
A Large Language Model (LLM) is a deep learning model trained on massive text datasets that can understand, generate, and reason about human language across a wide range of tasks.
