researchnokia2026-09-23
Nokia Applied Research open-sources AnyJev, an Apache-2.0 library for LLM-based typed decision models
Released under Apache 2.0 on September 23, AnyJev converts open-weight LLMs into typed decision models without further training. It uses probability distributions, permutation scoring, and calibration to improve robustness to option-order and label biases; Transformers and vLLM backends are supported.
诺基亚应用研究 9 月 23 日以 Apache 2.0 协议开源 AnyJev,一个 Python 库,用于将开放权重大语言模型转化为类型化决策模型而无需额外训练。库使用模型的概率分布、基于排列的评分与校准技术,使分类式决策对选项顺序与标签偏置更加鲁棒。
AnyJev 的设计目标是让开放权重 LLM 在「需要稳定决策」的工业场景中变得可信,例如路由、风险筛查、工具授权、自动化支持工单分流等场景。库内置 Transformers 与 vLLM 后端,便于在标准生产环境中集成。
诺基亚将其开源归入近期业界「让开放模型足够可靠」的更大方向:业界重点正在从单纯提高 LLM 生成质量,转向让开放模型在受限的生产决策场景中提供可校准的输出。AnyJev 的发布也是继 9 月 22 日 OpenAI 公布 AGMAI 数学顾问组、9 月 23 日多家公司发布开源/开放权重模型(MiMo-V2.6、Altar 1)之后,AI 研究与开源生态的又一具体进展。
nokiaanyjevopen-sourcedecision-models
Related News
10-22
research
MIT research finds 13 Claude agents engaged in unprompted play when left idle
10-22
research
METR and Redwood publish red-team report on OpenAI internal agent monitoring with novel bypasses
10-22
research
Microsoft MAI-DxO published in Science with 83% accuracy on complex cases versus 36% for physicians