CommunityJev Models & Reimplementations
Archai Jev Zagreus 0.4B Italian
An ultra-light Italian decision model: 400M Llama-based, a 465 MB Q8_0 GGUF, answering multiple choice with A–D tokens and booleans with TRUE/FALSE for edge and Rust use.
- Published by
- Community
- Author
- nickprock
- Last verified
- Oct 2026
- Capabilities
- Open sourceSelf-hostedRuns locally
Highlights
- It answers with tokens, not distributions: fixed A–D and TRUE/FALSE token ids are published on the card for constrained decoding.
- Trained with a joint cross-entropy and Brier loss; the card reports test accuracy 77.90% and Brier 0.2407.
- Built from seven native Italian sources, from hate-speech and ARC to sentiment and entailment.
- The Q8_0 GGUF is 465 MB and targets sub-10 ms CPU inference in edge and Rust environments.
Watch out
Self-reported only, with no independent evaluation. The card lists no context length, no options cap, no GPU requirements, and no limitations section, and the Jev branding is not backed by a Jev-format API.
Questions about this entry
Short answers drawn from the linked project or page.
Deterministic answer tokens: A, B, C, or D for multiple choice and TRUE or FALSE for booleans, with ids published (TRUE = 21260, FALSE = 31451). Parse those rather than free text, and note it returns no probability distribution.
The training objective combines cross-entropy with a Brier score term, which the card credits for keeping confidences calibrated. The only evidence shown is its own test Brier of 0.2407 across three epochs, with validation accuracy rising 77.04% to 82.18%.
It ships as a single Q8_0 GGUF of 465 MB, so it loads straight into RAM. The card lists llama.cpp, Ollama, vLLM, LM Studio, Jan and Docker Model Runner, for example `ollama run hf.co/nickprock/archai-jev-zagreus-0.4b-ita` or `llama serve -hf nickprock/archai-jev-zagreus-0.4b-ita`. It targets sub-10 ms CPU inference in edge and Rust environments.
It is a QLoRA fine-tune (r=16, alpha=32) of mii-llm/zagreus-0.4B-ita, a 400M-parameter model on the LlamaForCausalLM architecture. The developer is nickprock and the model is released under the apache-2.0 license.
A consolidated mixture of seven native Italian benchmark sources, including EvalITA, HateCheck-IT, RiTA-NLP and STS-B, spanning safety and moderation tasks such as hate speech detection through reasoning, sentiment and entailment.
The card publishes the target mappings for constrained decoding: multiple choice uses A=32, B=33, C=34, D=35, and boolean verification uses TRUE=21260 and FALSE=31451. Parsing those fixed ids is what produces the deterministic answer tokens.
No. The card lists no context length, no options cap, no GPU requirements and no limitations section, and the evaluation numbers shown are self-reported only with the model returning tokens rather than a probability distribution.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
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 NewsHacker News: A one-token logprob wrapper that also classifies webcam frames.
allanrbo151 pointsA single function Jev-like wrapper for LLMs, including vision modelsRead the thread on Hacker NewsMore Jev resources like this
1.3kGitHub stars
A local daemon that pulls open decision models and serves them on TypeSafe's /v1/systemone, so a Jev client switches hosts with TYPESAFE_BASE_URL.
Jev Models & Reimplementations#community#rust#open-models
Community
ollaya-dev/ollayaRust
ollaya-dev/ollayaRust877GitHub stars
A local server that reads next-token probabilities from a fine-tuned Spark model and returns Choice, Score, and Noul answers with zero generated tokens.
Jev Models & Reimplementations#community#python#open-models
Community
Rizzo-AI-Academy/rizzo-flowPython
Rizzo-AI-Academy/rizzo-flowPython58GitHub stars
A local /v1/systemone server that reads label probabilities from a GGUF chat model, with prompt repetition and a confidence shrink toward uniform.
Jev Models & Reimplementations#community#python#open-models
Community
Mushroom-Systems/lichenPython
Mushroom-Systems/lichenPythonThe Hugging Face hub for Laya's open System 1 checkpoints: the 421M English model at the repo root, a 322M multilingual build in a subfolder, and the typed-decisions fine-tune, all Apache-2.0.
Jev Models & Reimplementations#community#open-models#local
Community
The fine-tuned Laya checkpoint for four typed-decision workflows: 0.766 accuracy on 2,000 held-out decisions, above Jev 1.13.0's published 0.727 and the teacher self-agreement ceiling it trained past.
Jev Models & Reimplementations#community#open-models#local
Community
Supersonic Labs' 144.3M-parameter decision model on mmBERT-small: choice, score, and noul answers from supplied options in one pass, 73.15% on 2,000 typed decisions, Apache-2.0.
Jev Models & Reimplementations#community#open-models#local
Community
From the Jev guides
Original write-ups on what this entry is for.
Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.
Jev Tools & CLIs#open-models#local#onnx
Read article
Interface, size, license, claims — in that order. A selection guide across the system-one catalog, from a 706K form scorer to 27B Jev students.
Jev Practices & Patterns#open-models#evaluation#benchmarks
Read article
Hundreds of repositories carry the system-one tag, but the ecosystem is a dozen ideas in different sizes. A tour of the families, the licenses, and what a download does not include.
Jev Models & Reimplementations#open-models#hugging-face#local
Read article
From the Jev community
Posts from builders shipping with Jev right now.