LLM (Large Language Model)

Definition

A large language model is a transformer-based neural network, typically with billions of parameters, that has been pre-trained on broad text corpora. LLMs predict the next token from the input context and can therefore answer questions, summarize texts, write code, or invoke tools. Well-known examples include GPT, Claude, Gemini, Mistral, and DeepSeek. Quality, latency, cost, and privacy terms differ significantly between models and operators.

How Swiss Knowledge Hub uses this term

Swiss Knowledge Hub ships with deepseek-v3-0324, kimi-k2-5-instruct, and mistral-medium-2505 as default models. Through BYOK and depending on what a tenant has enabled, customers can additionally connect OpenAI, Azure OpenAI, Anthropic, Google, Mistral, DeepSeek, Azure DeepSeek, Azure AI Foundry, and OpenAI-compatible custom endpoints; in that case, data processing follows the terms of the chosen provider.

Related terms

• RAG (Retrieval-Augmented Generation)
• Context window
• Tokenization
• BYOK (Bring Your Own Key)
• Hallucination

Sources

1. Wikipedia: Large language model — https://en.wikipedia.org/wiki/Large_language_model