Autoregressive Model
The short answer · 15 words
In short
A model that predicts a next value or element using pre-existing or previously generated values.
Flow-alerta.cab (2026): Autoregressive Model. https://flow-alerta.cab/en/glossar/autoregressive-model. Editorially reviewed on: 09 October 2026.
How much detail would you like?
Practical
Basic term for time series and many generative models. It explains why outputs is gradual and why earlier outputs affects later predictions.
Example scenario: A language model generates a text token for token, with each new token depending on the previous context.
For forecasting or generation projects, check which context length, error propagation, latency and evaluation criteria are relevant.
At a glance
- Core
- Autoregressive models describe a quantity based on previous values. In generative AI, the principle is often used to gradually continue a sequence, for example by predicting a next token from the previous tokens. Statistical AR models and autoregressive language models use the same basic principle but are not the same model.
- Use
- Basic term for time series and many generative models. It explains why outputs is gradual and why earlier outputs affects later predictions.
- Example
- Example scenario: A language model generates a text token for token, with each new token depending on the previous context.
- Limits
- Call each sequential model autoregressive; equate statistical AR model with LLM; overlook dependence on previous model predictions; ignore error propagation.
In context
Questions about this term
Is every language model autoregressive?
No.
Are AR time series models and LLMs the same?
No.
Why can mistakes continue?
Because later steps may depend on previous values or generated elements.
All details and background
Autoregressive models describe a quantity based on previous values. In generative AI, the principle is often used to gradually continue a sequence, for example by predicting a next token from the previous tokens. Statistical AR models and autoregressive language models use the same basic principle but are not the same model.
ELI5 – Explained simply
The model looks at what was there before and estimates the next step.
Why it matters in practice
Basic term for time series and many generative models. It explains why outputs is gradual and why earlier outputs affects later predictions.
Example
Example scenario: A language model generates a text token for token, with each new token depending on the previous context.
Pitfalls and limits
- Call each sequential model autoregressive; equate statistical AR model with LLM; overlook dependence on previous model predictions; ignore error propagation.
Origin & history
Autoregression is used both in statistical time series models and in the stepwise generation of sequences. The concrete models and their representations differ.
Use in mid-sized companies
- For forecasting or generation projects, check which context length, error propagation, latency and evaluation criteria are relevant.
Related terms
- ARIMA model
- Tokens
- Time series
Matching Flow-alerta service
AI opportunity analysis & roadmapAI consulting: AI potential analysis & roadmap on tasks, data and prerequisites of a possible first AI use case.Matching tool / use case
View use casesUse Case: Token-wise text generation or prediction of a time series from past values.Sources
- developers.google.comGoogle for Developers, Machine Learning Glossary – auto-regressive model
Quellenprüfung: 09.10.2026. Definition und dokumentierte Einordnung abgeglichen; Beispiele beschreiben mögliche Anwendungen.
Editorially reviewed on: