Luce: train and calibrate a decision model for your task

A task-specific pipeline generates or labels data with an LLM teacher, trains a LoRA and decision head, then evaluates and serves Choice, Score and Noul probabilities. A public support-triage replay shows saved model outputs and review routing.

Open demo ↗

Notes

The author reports 91.1% accuracy and ECE 0.022 on 2,964 rule-labelled question answers from 988 tickets after training Qwen3-4B-Base on 3,000 LLM-written tickets for two epochs.

source-backed — Public repository, docs or live artifact. About this label

Destinations

Filed under

• #local
• #training
• #calibration
• #decision-making

In collections

• Benchmarks & independent tests →
• Local & replica models →