Reasoning model
A reasoning model is a language model trained to produce extended internal deliberation before answering, spending more computation on hard problems.
Reasoning models change prompting practice: instructions to "think step by step" are redundant and can interfere, since the deliberation already happens. What still helps is a clear statement of the goal, the constraints, and how the answer will be judged.
Related terms
- Chain-of-thought prompting
- Chain-of-thought prompting is instructing a model to work through its reasoning step by step before giving an answer, rather than answering immediately.
- Temperature
- Temperature is a setting between 0 and roughly 2 that controls how much randomness a model uses when choosing each next token.
More on retrieval and agents
See the full glossary, read the guides, or put it into practice in the prompt builder.