awesome-jev-security
Curated list of papers on the security, robustness and safety of Jev and RLCD decision models, from attacks to deployment.
This repository is mostly documentation or write-up rather than runnable code.
Overview
Awesome Jev Security tracks research on the security, robustness and safety of Jev and of RLCD (Reinforcement Learning for Calibrated Decisions) models more broadly. It notes that typed outputs rule out malformed responses but change the attack surface: attackers can still push the decision distribution or confidence scores.
Sections cover attacks and adversarial robustness, Jev as a security or safety detector, Jev in security applications, reliability and safe deployment, background and related projects. Early entries include a broad study of security, privacy and dual use, an adversarial benchmark where a single appended opinion flips a share of decisions, a paper on natural context flipping decisions, and work on prompt injection against typed decisions.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Other | How can Jev's decisions be attacked, and how should it be deployed safely?(inferred) | — |
How it works
- Track new papers
- Group by attack, detection, application or deployment
- Summarize key findings
Strengths
- Security-specific focus
- Short summaries of key papers
Caveats
- Early, preprint-heavy literature
At a glance
- Type
- Article
- Status
- research
- Has code
- no
- Has tests
- no
Use cases and integrations
How curated lists describe it
Curated list of papers on the security, robustness, and safety of TypeSafe Jev / System One models.
hellogumbo/awesome-jev
Frequently asked questions
Can Jev be attacked if it never writes text?
Yes, attackers can still shift decision probabilities and confidence, as the listed papers show.
What is RLCD?
Reinforcement Learning for Calibrated Decisions, the training approach associated with Jev-like models.
How current is it?
It tracks papers since the 2026-09-15 launch.
Related projects
Bilingual catalogue of Jev papers, software projects and public materials, with evidence levels, published as a static site.
AppproductionIndependent, source-linked research on Jev with about 948 rubric-scored public repositories, recurring patterns and verified findings.
ProjectresearchAwesome JEV gallery: curated papers, open reproductions, benchmarks and independent evaluations for Jev and System One models.
ArticleproductionMarkdownList of research papers on Jev and the open models built in its shape, each checked against its arXiv page.
ArticleproductionEvidence-aware map of 137 research items on Jev and typed probabilistic decisions, with evidence tiers and an interactive atlas.
ArticleproductionLibrary of 389 papers, projects and documents on Jev's foundations, trustworthiness and security uses, with a daily papers feed.
Articleproduction