Rubric-based eval harness cheap enough to run on every PR, powered by typesafe-ai/jev
Research and benchmarks — page 8
Benchmarks, papers and independent reproductions. Read these when you need to understand behaviour rather than adopt a library.
537 active projects in this collection. All collections →
Fast and cheap agent evals. jev as judge.
TypeSafe System One models (Jev) for RubyLLM: typed judgments, evaluations and reranking.
3
Measure when to use Jev and other models on your data, then route accordingly.
3
An eval-first MCP server for TypeSafe's Jev, a System One model that returns typed judgments (noul, choice, score) with probabilities instead of generated text.
Reproducible Jev Ultrafast research-browser eval harness + field note (QC’d cases, suite runner, report generator). Not investment advice.
Benchmarking TypeSafe's Jev decision model as a cost-efficient LLM router on RouterArena
Measures how well TypeSafe's RLCD-Jev model spots real secret credentials in file snippets
Zero-shot spam filtering with TypeSafe Jev Noul questions, compared with TF-IDF baselines
No project description available.
Survey of the open-source ecosystem around TypeSafe Jev: use-case classification of 16,128 GitHub repositories, verified integrations, replicas, and evaluations (English and Simplified Chinese)
A reproducible, self-hosted benchmark for comparing JEV and LLM decision-making on your own datasets, models, and evaluation policies.
JeVerifier: cheap Jev (TypeSafe) checks that keep code maintainable and docs consistent, plus context retrieval — modest token savings
Typed, probabilistic decisions (Choice / Score / Noul) for System One tasks, powered by TypeSafe Jev, with OpenAI and Anthropic as alternative engines
Benchmark of TypeSafe's Jev as a malicious agent-skill detector on MalSkillBench, with a verify-and-escalate cascade
Creating system one model just like jev. with different experimentation
A fast System One decision model (Jev or local Laya) picks which LLM answers each prompt: OpenAI-compatible Go gateway + decision benchmark
Ne yazarsanız yazın, beş kelimeden biriyle cevap veren sohbet botu. Kelimeyi TypeSafe AI'ın Jev evaluation modeli seçer.
Deep research crawl where Jev (a System One model) makes every per-page decision and an LLM only plans and writes
TypeSafe's Jev and four fast LLMs added to Stanford MedHELM's MedHallu results: harness, preregistered run plans and every run file.
Grade the agent sessions already on your disk. Claude Code, Codex, opencode, Gemini CLI and Antigravity, scored with Jev for cents.
Windows port of Laya typed-decision AI (ONNX Runtime + DirectML) — Core ML/Apple Neural Engine alternative with first-class Arabic support. No text generation, no hallucination, runs on any DX12 GPU.
Metask-Jev: calibrated typed-decision models (Jev-class). Single forward pass, candidate-logit readout. metask-jev-4b beats Bespoke Nimble-9B and Jev on JevBench.
Offline $0 decision layer for coding agents: Choice/Score/Noul primitives, BELKI confidence gatekeeper, ultra-planning, red-teaming, research and RLVR self-improvement -- as an MCP server + CLI + Claude Code skill.