Organization repository on GitHub
- Forks
- 4
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Demos & Experiments
S1-CED Playground
A browser playground for S1-CED, a small decision model you give a piece of text plus typed questions — yes/no, which-of, or a 0-3 score — and it answers with a probability per option and whether it is confident enough to act or a human should check. Runs on CPU; no GPU or account needed.
- Stars
- 6
- Language
- Python
- License
- APACHE-2.0
- Reading time
- 2 min
- Published by
- Community
- Author
- ai-yatra
- Last verified
- Oct 2026
Used inJev for Software & DevTools
Highlights
- Runs on your own computer: the README requires no GPU and no account, just uv and Python.
- Question types shown: yes/no safety checks, choice between teams, and 0-3 scores like customer frustration.
- Every answer shows the probability of each option plus whether the model is confident enough to act or a human should check.
- Example inputs include a support ticket, a resume and a proposed agent action.
- Apache-2.0 per repository metadata; created 2026-10-15.
Watch out
README reviewed 2026-10-17; the digest truncates the Step 2 launch command mid-flag (uvx --torch-backend cpu) and the repository's GitHub description is empty, so the exact run command, model size and licence text could not be verified. The README reports no accuracy numbers for S1-CED.
Questions about this entry
Short answers drawn from the linked project or page.
A small local playground for S1-CED, a model that answers questions about a piece of text with probabilities instead of just answers. You give it text and typed questions, and it shows the distribution over answers for each one.
Three shapes are shown in the README: yes/no, such as whether an action is safe to run without approval; choice, such as which team should own a ticket; and score, such as how frustrated a customer is from 0 to 3.
No. The README says it runs on your own computer with no GPU and no account needed. Setup starts by installing uv and then launching the playground with a uvx command whose tail the digest truncates.
Install uv with the documented one-line installer for macOS/Linux or Windows, open a new terminal so uv is found, then run the Step 2 launch command from the README. The digest cuts that command off after uvx --torch-backend cpu, so read the exact flags from the repository.
The playground shows the probability of every option and states whether the model is confident enough to act on, or whether a human should check — the confidence-gating pattern the README builds the interface around.
A support ticket, a resume, and a proposed agent action. For the ticket it asks which team should own it; for actions it asks whether the step is safe without human approval; scores appear as a 0-to-3 scale for frustration.
The repository belongs to AI-Yatra, created 2026-10-15 and pushed the next day, with 6 stars and 3 forks. GitHub metadata lists Apache-2.0, though the digest shows no licence section in the README text.
More Jev resources like this
An interactive ZeroGPU demo of AutoTrust's JEV-9B: System 1 answers noul, choice, and score questions in one prefill pass with calibrated probabilities, while an Escalate tab hands low-confidence cases to the untouched Qwen3.5-9B base.
Jev Demos & ExperimentsPlayground#community#playground#demo
Community
A ZeroGPU demo of AutoTrust's JEV-27B-VL: System 1 returns calibrated noul, choice, and score probabilities from one prefill pass over text and images, with System 2 on the untouched Qwen3.8-27B for escalation.
Jev Demos & ExperimentsPlayground#community#playground#demo
Community
A frozen decision server that serves Gemma 4 31B — quantized to FP8 at startup — in TypeSafe's /v1/systemone format, reading the next-token distribution instead of generating tokens. It needs one FP8 NVIDIA GPU with 80 GB and has no authentication.
Jev Repos & SDKs#community#github#python
Community
krishna-gogineni-765/deck31bPython
krishna-gogineni-765/deck31bPythonA workshop of small, inspectable Python experiments for TypeSafe System One models: ten critical bays, from fraud screening to emergency dispatch, each giving Jev a realistic state while deterministic policy keeps authority over the next action.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
jgalego/jevs-garagePython
jgalego/jevs-garagePythonA standalone proof of concept showing Jev as a structured decision layer for support-ticket routing: fixed-option routing, 1–5 urgency scoring and probabilities for sensitive data, with a deterministic confidence and safety policy deciding auto-route versus human review.
Jev Demos & Experiments#community#github#streamlit
A three-lane demo of pngwn's System One 4B scorer: one prefill plus a batched forward for every option branch, a naive per-option equivalence check, and a Qwen3.5-4B instruct baseline generating JSON.
Jev Demos & ExperimentsPlayground#community#playground#demo
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.