Chinese-Jev: Bringing System One Model to Chinese-Language Tasks

Hugging Face Daily Papers Papers
chinese-language system-one-model decision-making fine-tuning domain-specific on-device-deployment

Summary

This paper introduces Chinese-Jev, a System One model designed to improve decision-making accuracy and efficiency for Chinese-language tasks, with domain-specific fine-tuning and a new benchmark, CJ-Bench, for evaluation.

System One models such as Jev offer an efficient alternative to generative language models for tasks that require decisions rather than open-ended responses. However, existing Jev models exhibit limited Chinese-language decision accuracy, restricting their utility in both general and specialized settings. In this paper, we introduce Chinese-Jev, a System One model that addresses this gap through a unified data processing and training pipeline. Our data processing protocol converts heterogeneous Chinese-language annotations into probability targets over candidate options, enabling a shared training formulation across domains and question formats. To enable efficient inference, Chinese-Jev adopts a lightweight encoder-only backbone for text encoding and learns to score candidate answers through decision-oriented training. To address the misalignment between the pre-training distribution and downstream Chinese-language scenarios, we first train the model on a general-purpose corpus of 10 million examples, then fine-tune it separately for the medical, legal, and financial domains. To evaluate decision accuracy and calibration in both general and domain-specific Chinese-language settings, we introduce Chinese-Jev Bench (CJ-Bench). After first-stage pre-training, Chinese-Jev exceeds the accuracy of the closed-source Jev model by 1.24% on general-domain tasks while achieving a 20.3x speedup. Subsequent domain-specific fine-tuning yields a 4.0% accuracy improvement over Jev in medicine and achieves 92% of Jev's average accuracy across specialized domains, with a 17x speedup and an average latency of only 15 ms per example. We further demonstrate on-device deployment of an INT8-quantized model on mobile devices, achieving an inference latency of approximately 1.0 second per decision. The project is available at https://gulucaptain.github.io/Chinese-Jev/.
Original Article
View Cached Full Text

Cached at: 09/30/26, 08:22 AM

Paper page - Chinese-Jev: Bringing System One Model to Chinese-Language Tasks

Source: https://huggingface.co/papers/2609.36965

Abstract

SystemOnemodelssuchasJevofferanefficientalternativetogenerativelanguagemodelsfortasksthatrequiredecisionsratherthanopen-endedresponses.However,existingJevmodelsexhibitlimitedChinese-languagedecisionaccuracy,restrictingtheirutilityinbothgeneralandspecializedsettings.Inthispaper,weintroduceChinese-Jev,aSystemOnemodelthataddressesthisgapthroughaunifieddataprocessingandtrainingpipeline.OurdataprocessingprotocolconvertsheterogeneousChinese-languageannotationsintoprobabilitytargetsovercandidateoptions,enablingasharedtrainingformulationacrossdomainsandquestionformats.Toenableefficientinference,Chinese-Jevadoptsalightweightencoder-onlybackbonefortextencodingandlearnstoscorecandidateanswersthroughdecision-orientedtraining.Toaddressthemisalignmentbetweenthepre-trainingdistributionanddownstreamChinese-languagescenarios,wefirsttrainthemodelonageneral-purposecorpusof10millionexamples,thenfine-tuneitseparatelyforthemedical,legal,andfinancialdomains.Toevaluatedecisionaccuracyandcalibrationinbothgeneralanddomain-specificChinese-languagesettings,weintroduceChinese-JevBench(CJ-Bench).Afterfirst-stagepre-training,Chinese-Jevexceedstheaccuracyoftheclosed-sourceJevmodelby1.24%ongeneral-domaintaskswhileachievinga20.3xspeedup.Subsequentdomain-specificfine-tuningyieldsa4.0%accuracyimprovementoverJevinmedicineandachieves92%ofJev’saverageaccuracyacrossspecializeddomains,witha17xspeedupandanaveragelatencyofonly15msperexample.Wefurtherdemonstrateon-devicedeploymentofanINT8-quantizedmodelonmobiledevices,achievinganinferencelatencyofapproximately1.0secondperdecision.Theprojectisavailableathttps://gulucaptain.github.io/Chinese-Jev/.

View arXiv pageView PDFProject pageGitHub1Add to collection

Get this paper in your agent:

hf papers read 2609\.36965

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.36965 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.36965 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.36965 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles