Three measured experiments on RAG hallucination: quote-checking, TypeSafe's Jev, and IBM's STAIR. 850+ graded questions, raw responses included.
仓库包含对TypeSafe AI Jev模型在RAG场景中的多项实测评估（幻觉抑制、重排序、路由、重试决策），提供详细报告与可复现数据，属于实质性测试研究。
JevalsJevals/jevals-dataIndependent benchmark data comparing TypeSafe AI's Jev (System One model) against LLMs on decision score, accuracy, calibration, cost, and latency for typed tasks.
ReallyArtificialReallyArtificial/stuntdoubleA zero-dependency Node proxy that shadows TypeSafe Jev with local models (Kev, Laya) on live traffic and produces a report on whether you can swap.
NearCaiNearCai/JevLightA CityFlow-based traffic signal control framework that evaluates TypeSafe Jev and OpenAI-compatible LLM controllers alongside rule-based and reinforcement-learning baselines through the same simulation pipeline.
