Jeff – Jev-compatible 0.8B decision models, trained at home, ~30 ms

Jeff is a new decision-making model that can be easily integrated into local code for fast, well-calibrated judgments between options. It uses a simple request format similar to Jev and does not require any generated text or parsing. The model is trained on synthetic data and can handle various types of questions, such as choice, yes/no, and scoring. With its small size and fast inference time (around 22-30 ms per decision), Jeff is well-suited for local applications where speed and calibration are important.

Brief written by urgent.news from Hacker News's own syndicated text. Machine-written — may contain errors; check the original before relying on it.