#AnyJev
Jiamu Zhang, Tianze Yang, Yucheng Shi, Evan Chen, Zixiang Nie, Kelly Wan, Liangjie Hong, Ninghao Liu, Liang Wu: AnyJev Technical Report https://arxiv.org/abs/2610.00831 https://arxiv.org/pdf/2610.00831 https://arxiv.org/html/2610.00831
October 2, 2026 at 6:44 AM
Nokia’s new AnyJev library wipes out LLM biases in decision models—no extra training needed. Think fine‑tuned performance with just a Python tweak. Curious how logits get smarter? Dive in! #AnyJev #LLMBiasFix #DecisionModels

🔗 aidailypost.com/news/nokias-...
September 23, 2026 at 7:21 AM
Nokia Open-Sources AnyJev: Training-Free Layer Calibrating Open LLMs In enterprise AI, the question isn’t only what a model can generate; it’s how reliably.... @thecosmicmeta.com

https://u2m.io/xj16Q7ML
Nokia Open-Sources AnyJev: Training-Free Layer Calibrating Open LLMs
In enterprise AI, the question isn’t only what a model can generate; it’s how reliably those outputs drive real decisions.
thecosmicmeta.com
September 23, 2026 at 4:17 PM
「判定を返すLLM」がOSSになった。NokiaのAnyJevは、学習なしで任意のオープンLLMを確率付きの判定モデルに変える層。文章は返さず、選択肢から1つを選ぶだけ。Apache-2.0でPyPIから入る。出力をJSONパースで整える苦労が、最初から無くなるのは正直ありがたい。Jevの系譜が一気に実用圏に来た気がする。#AnyJev
September 23, 2026 at 10:10 AM
Nokia open-sourct AnyJev: eine Schicht, die jedes offene LLM ohne Training zum Entscheidungsmodell macht – eine Option aus fester Menge statt freier Text. Spannend ist weniger die Lib als die Richtung: Schichten entscheiden über Produktionsreife.…

Mehr von mir: linkedin.com/in/maurice-putinas
September 23, 2026 at 3:00 PM
The Python library uses an adapter mechanism to fix prior and position biases in LLMs, enabling typed decision models.
Nokia ships AnyJev: open LLMs become calibrated deciders
The Python library uses an adapter mechanism to fix prior and position biases in LLMs, enabling typed decision models.
www.alextech.ai
September 23, 2026 at 1:58 PM