先例清单：ibl.ai 指出 CriticGPT 与 Prometheus 2 都在 2024 年就做"判断模型" → 写进优先权争议页 §4d，与 Laya 的指称分开记；本库判断 Jev 与它们的真实差异是"把判断做成带类型与概率的一等 API"
ibl.ai article noting CriticGPT and Prometheus 2 built judge models in 2024, cited as prior art for Jev's design.
This summary is based on how curated lists describe the project; its README was not available to review.
Overview
This entry points to an ibl.ai blog post about Jev as a judge model for evaluation rather than generation in enterprise settings. The wh000wh000/awesome-jev-live list records it as a precedent: the article points out that OpenAI's CriticGPT and Prometheus 2 were already building judgment models in 2024. The list keeps this separate from LAYA's own priority claims and concludes that Jev's real difference from those systems is turning judgment into a first-class API with types and probabilities.
It is relevant for readers tracing where decision models came from and how Jev differs from earlier LLM-as-judge work. The article itself was not fetched, so its specific arguments are summarised only as the list describes them.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Other | How does Jev compare to earlier judge models such as CriticGPT and Prometheus 2?(inferred) | — |
How it works
- Review prior judge models from 2024
- Compare them to Jev's typed, probabilistic API
- Record the distinction in a priority-dispute page
Strengths
- Gives historical context for decision models
- Clarifies that typed outputs with probabilities are the novel part
Caveats
- Vendor blog post; not peer-reviewed
- Only the list's paraphrase was available
At a glance
- Type
- Article
- Status
- discussion
- Has code
- no
Use cases and integrations
How curated lists describe it
Precedent list: ibl.ai points out that CriticGPT and Prometheus 2 were…
wh000wh000/awesome-jev-live
Frequently asked questions
What precedent does the article raise?
That CriticGPT and Prometheus 2 were already judge models in 2024.
What is Jev's difference, per the list?
Making judgment a first-class API with typed answers and probabilities.
Is this related to the LAYA priority debate?
The list files it in the same dispute page but keeps it separate from LAYA's claims.
Related projects
RoboKrunch's curated Jev resource list with official links, news and its own measured experiments marked separately.
ProjectproductionIndependent, source-linked research on Jev with about 948 rubric-scored public repositories, recurring patterns and verified findings.
ProjectresearchLong-form analysis of what Jev is, why it is fast, how much of the hype holds up, and how hard it is to reproduce.
ProjectdiscussionJapanese research deck, article and source ledger tracking TypeSafe's Jev and System One models over time.
ProjectdiscussionMarkdownUnofficial Jev information hub site with access guide, pricing comparison, a BYOK playground and an auto-updated news timeline.
AppproductionPythonChinese-run aggregator of X long posts and demo videos about Jev, keeping original links and authors rather than rehosting.
Serviceproduction