OpenJev Verdict 2.0

# openJev-verdict-2.0: Non-Autoregressive System 1 Decision Engine

[](https://huggingface.co/heman10x/openJev-verdict-2.0)
[](https://github.com/Heman10x-NGU/openJev-verdict-2.0)
[](#the-openjev-verdict-20-benchmark-breakthrough)
[](#dual-channel-calibration)
[](#single-pass-efficiency)
[](#in-browser-webgpu-engine)
[](LICENSE)

---

## Two models in this project

This repository contains code and references for two distinct models:

1. Verdict (the 151M model evaluated on JevBench): General-purpose decision model based on ModernBERT-base and GLiClass. The public checkpoint is hosted at [heman10x/rlcd-modernbert-151m](https://huggingface.co/heman10x/rlcd-modernbert-151m), now updated to v1.4 via inference engine fixes.
2. Verdict 2.0: Specialized architecture for typed software workflows evaluated on `LocalLLaMA/typed-decisions`. The model weights are currently tracked via Git LFS pointers in `artifacts/verdict2-base/model.pt`. The benchmark numbers reported in the breakthrough section below were produced directly by `verdict2/evaluate.py` against its audited test receipt (`reports/verdict2_base_test.json`).

---

## What changed in the inference engine

These are inference fixes, not a retrain. The weights are byte-identical to the published checkpoint. Measured on the 231 public JevBench tasks.

<p align="center">
  <img src="assets/v1.4/benchmark-leaderboard-chart.png" alt="JevBench Score Leaderboard: openJev Verdict v1.4" width="100%">
</p>

<p align="center">
  <img src="assets/v1.4/benchmark-leaderboard-table.png" alt="JevBench Detailed Evaluation Table: openJev Verdict v1.4" width="100%">
</p>

The update addresses three defects in the inference engine:

1. Calibrator auto-loading and removal of the 5-option scope restriction: The engine previously failed to loa