User repository on GitHub
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Tools & CLIs
JevSoup: System-One Routing for Training-Free LoRA Composition (arXiv 2609.30922)
JevSoup is a training-free framework that separates System One expert routing from System Two execution: Jev picks two LoRA experts through structured probabilities and their updates are merged with orthogonal projection. Across 14 PorTAL tasks it gains up to 1.19% task-macro accuracy over external baselines.
- Stars
- 1
- Language
- Python
- Format
- Article
- Published by
- Community
- Last verified
- Oct 2026
Highlights
- Uses only the input and expert descriptions; Jev selects two experts through structured probabilities.
- Retains the leading expert's update and projects the second onto the orthogonal complement of the first's row space.
- Combines the two updates with equal weights after the projection.
- Across 14 PorTAL tasks and three Qwen3 scales, gains up to 1.19% task-macro and 1.21% sample-micro accuracy.
Watch out
arXiv preprint dated September 25, 2026 (cs.AI), five pages, listed as under review, by Xiuying Wang and co-authors. The gains are self-reported against the strongest evaluated external baselines on 14 PorTAL tasks with three Qwen3 scales, and the abstract gives no per-task breakdown.
Questions about this entry
Short answers drawn from the linked project or page.
It separates System One expert routing from System Two execution: Jev handles expert selection, while the chosen experts' LoRA updates are merged and executed without the router being trained.
No. JevSoup is training-free and uses only the input and expert descriptions, unlike routing approaches that may require auxiliary data, additional training, or autoregressive decoding.
JevSoup retains the leading expert's update, projects the second onto the orthogonal complement of the first update's row space, and combines them with equal weights.
Across 14 PorTAL tasks and three Qwen3 scales, JevSoup reports absolute gains of up to 1.19% in task-macro and 1.21% in sample-micro accuracy over the strongest evaluated external baselines.
Jev selects two experts through structured probabilities based only on the input and the expert descriptions, which is what makes the composition training-free.
Yes. The abstract states the code is available at a public GitHub repository, making the training-free routing and projection scheme inspectable.
It is a five-page preprint dated September 25, 2026, listed as under review, so the reported gains have not been through peer review and cover only the evaluated PorTAL tasks.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
Facebook: Intake Router keeps diagnosis out of the answer space at the front desk.
FacebookClinicnotes
YouTube: Jev explained for routing and judging, with a comparison against LLMs.
Hacker News: Vercel on using Jev for defined decisions and GPT-6 Astra when the job also has to write or use tools.
flashbrew4 pointsTypeSafe AI Jev vs. GPT-6 AstraRead the thread on Hacker NewsHacker News: ASTGL replay chose local Laya routing over hosted Jev on one Mac Studio setup.
Jmeg8r15 pointsDoes an open-weight decision model beat a hosted one? Jev vs. LayaRead the thread on Hacker NewsMore Jev resources like this
A semantic database system that answers filters, joins, classification, and ranking with fast typed decision models and escalates uncertain cases to LLMs. JEVDB-Flash posts the lowest latency on all 21 SemBench queries, and Semantic Bloom Filters remove 87.4% of join candidates on Shelob.
Jev Tools & CLIsArticle#community#research#arxiv
Community
Jev-Mem is an agentic memory architecture that gives a System-One control plane charge of memory typing, query routing, retrieval budgets, graph traversal, candidate scoring, and adaptive stopping, invoking System Two only for complex reasoning and synthesis. On LoCoMo it scores 0.777.
Jev Tools & CLIsArticle#community#research#arxiv
Community
A system that joins coded and narrative views of 5,601,890 Texas crashes from 2017 to 2025 into population estimates: Kumo Tabular reads coded records, Jev reads narrative samples, and human judgments recalibrate the probabilities. Phone use appears in 15,074 injury crashes in narratives versus 7,340 in coded fields.
Jev Tools & CLIsArticle#community#research#arxiv
Community
pozapas/kumo-jev-crash-recordsPython
pozapas/kumo-jev-crash-recordsPythonAn integration of Jev's decision API into edge service orchestration that extracts four to eight bounded intent fields per request. Across 33 test conditions Jev cuts median decision latency by 22.7-64.5% versus the fastest LLM, and on a live admission path it keeps 0.91-0.95 of requests exact and on time.
Jev Tools & CLIsArticle#community#research#arxiv
Community
1.9kGitHub stars
A Go-based team workspace with optional Jev AI for agent auto-routing, task auto-fill, auto-assign, and automation condition nodes.
Jev Tools & CLIs#community#go#agents
Community
Paca-AI/pacaGo
Paca-AI/pacaGo1GitHub stars
KITE queries a typed behavioral kernel once per unique state, then runs populations of any size from the table, reserving an expensive flagship model for sparse paired anchors. On Epstein experiments with 9,070 participants, anchors covering 1.7% of states cut effect error by 41%.
Jev Tools & CLIsArticle#community#research#arxiv
From the Jev guides
Original write-ups on what this entry is for.
The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.
Jev Tools & CLIs#gateway#vercel#cloudflare
Read article
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
Jev labels a record that already exists. Code moves, flags, or holds it. How the document, mail, and transfer tools keep the action out of the model.
Jev Tools & CLIs#classification#email#security
Read article
From the Jev community
Posts from builders shipping with Jev right now.