Jev directory
CommunityJev Models & Reimplementations
V Jev 1.4
v-jev 1.4 is a Jev-style decision model on Hugging Face in 0.8B and 4B sizes, built on Qwen3.5 backbones. It returns probabilities for typed questions in one forward pass and adds value extraction and image decisions.
- Reading time
- 2 min
- Published by
- Community
- Added
- Nov 2026
Highlights
- Turns a state and typed questions into a probability for every allowed option in one forward pass.
- Version 1.4 adds value extraction for dates, amounts, ids, names, addresses, and urls, plus image decisions.
- Released in 0.8B and 4B sizes on Qwen3.5-0.8B and Qwen3.5-4B backbones, each in its own folder.
- On typed-decisions, 4B scored 0.638 accuracy and 0.272 KL from gold; 0.8B scored 0.481 and 0.372.
- Research release: one training run and one seed per size, only English tested, license apache-2.0.
Questions about this entry
Short answers drawn from the linked project or page.
Install transformers and accelerate, then load with AutoModel.from_pretrained("haohan84/v-jev1.4", device_map="auto"). The repo provides Safetensors weights and a Transformers model card.
No. It turns a state and a set of typed questions into decisions in a single forward pass: a probability for every allowed option of every question. Nothing is generated and nothing needs parsing.
Two sizes: 0.8B on Qwen3.5-0.8B and 4B on Qwen3.5-4B. Each size is in its own folder and post-trained the same way from its own backbone.
The model card lists License: apache-2.0. It is a research release with one training run and one seed per size.
Only English was tested. The model card does not claim support for other languages.
For a text/image request: 1.5/1.8 GB for the 0.8B size and 8.0/9.1 GB for the 4B size. Weight files are 1.7 GB and 9.1 GB respectively.
It is a research release with one training run and one seed per size, tested only in English. Decisions are trained only on the backbone's own answers, so they show what the backbone itself thinks rather than adding knowledge.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Standard One SH (StandardThinking)
LinkedIn: What multimodal decision models actually change, and what they do not.
LinkedInJevDirectory.org
Hacker News: Decision Index 0.2.1 scores open reproductions of Jev on a frozen suite.
jackbravo4 pointsDecision Index 0.2.1 – 70 open reproductions of TypeSafe Jev's Decision ModelRead the thread on Hacker NewsHacker News: Pointer to a DiffusionGemma plus vLLM tutorial for a Jev-shaped classifier.
1 pointHow to build a Jev-style classifier with DiffusionGemma and vLLMRead the thread on Hacker NewsMore Jev resources like this
A 7.7M-parameter decision model on Hugging Face that reads text or speech and returns one probability per option for choice, score, and noul questions. Speech runs through femto-asr, giving a 10.1M-parameter spoken pipeline.
Jev Models & Reimplementations#open-models#classification#local
Community
arXiv paper presenting EDU-Jev, a family of single-pass System One decision models at 0.8B, 2B and 4B parameters that return a probability distribution over predefined education answer options without generating any text.
Jev Models & ReimplementationsArticle#arxiv#open-models#classification
Community
MLX builds of Jev-Style-2B-Decision-v4 for Apple silicon: bf16 (3.76 GB) and 8-bit (2.00 GB) folders for text and JSON states, run through one shared jev_style_decision_mlx.py runtime under Apache-2.0.
Jev Models & Reimplementations#open-models#calibration#local
Community
A 26M-class decision model on Hugging Face, fine-tuned from q-project/N-Q-26M-Chat. It ships a custom NQ4ForDecision class for text classification and feature extraction, is tagged calibrated and decision-model, and publishes Apache-2.0 weights with Russian and English support.
Jev Models & Reimplementations#open-models#calibration#classification
Community
Fine-tuned UltraBERT-380M decision model on Hugging Face that scores a state, question, and option list in one forward pass. Uses sigmoid of yes minus no normalized over options; 72.8% dev and 71.9% test accuracy across 13 datasets, with an evidence head documented on the card.
Jev Models & Reimplementations#open-models#arxiv#calibration
Community
A Hugging Face checkpoint that fine-tunes UltraBERT-380M into a Jev-style decision model: it reads a state, a question and a list of options and picks one in a single forward pass. Reports 72.6% mean accuracy over 13 datasets on the test split.
Jev Models & Reimplementations#open-models#classification#local
Community
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.