Jev introduces a new shape of LLM - System One, aka Decision Models
How-To How to actually use this
What changed: TypeSafe AI released Jev, a "System One" or "decision model" LLM that outputs floating point numbers for categories or yes/no instead of text.
How to use it:
- Identify a task requiring a numerical score or binary classification rather than generated text.
- Send your text prompt to the Jev API as you would to a standard LLM.
- Parse the returned floating point values to map to your specific categories or decision thresholds.
- Integrate the numerical outputs directly into your automation or scoring logic.
Good for: developers building classification or scoring systems without text-parsing overhead.
Last week TypeSafe AI unveiled Jev , their first example of a new category of model that they are calling "System One models" (I'm with Maggie Appleton, I think "decision models" is a better name for these). Jev is an interesting variant on the usual LLM format: it still accepts text inputs, but instead of text output it returns floating point numbers corresponding to categories, yes/no…
Read original article on Simon Willison's Weblog →More in Model Releases
