Hallucination (AI)

Language models

A hallucination is an incorrect, approximate or invented answer that a language model delivers with confidence. It stems from the probabilistic workings of LLMs: the model aims above all to produce linguistically coherent text, even without reliable information on the subject.

In practice at Gensai

Gensai reduces hallucinations by grounding answers in verified data through RAG, structuring prompts and keeping human validation on critical uses.

Same theme

LLM (Large Language Model)SLM (Small Language Model)DLLM (Diffusion Large Language Model)Prompt chaining Pre-trained model Foundation model Generative AI Transformer Token Context window Prompt Prompt engineering Context (AI)Memory (AI)Fine-tuning

Hallucination (AI)

In practice at Gensai

Go deeper

Same theme