PAST-TIDE: Prototype-Anchored Statement Tuning with Topic-Invariant Normalization for Stance Detection

Hugging Face Daily Papers Papers

Summary

PAST-TIDE is a stance detection system for the StanceNakba Shared Task, using statement tuning with cloze-style masked language modeling, prototypical contrastive learning, and topic-conditional layer normalization for cross-topic Arabic stance detection, achieving macro-F1 scores of 0.75 and 0.74 on subtasks A and B.

We introduce PAST-TIDE, our stance detection system addressing both subtasks of the StanceNakba Shared Task at NakbaNLP@LREC-COLING 2026. The main idea is statement tuning. We redefine stance as cloze-style masked language modeling (MLM), letting a verbalizer map label words to stance categories through the pre-trained MLM head rather than appending a randomly initialized classification head. We complement this with prototypical contrastive learning, which uses learnable class prototypes for batch-size independent contrastive training, and topic-conditional layer normalization for cross-topic Arabic stance detection. PAST-TIDE achieves macro-F1 scores of 0.75 for Subtask A and 0.74 for Subtask B on the official leaderboard, indicating that minimal architectural additions to a pre-trained model can remain competitive in low-resource settings.
Original Article
View Cached Full Text

Cached at: 07/10/26, 10:12 PM

Paper page - PAST-TIDE: Prototype-Anchored Statement Tuning with Topic-Invariant Normalization for Stance Detection

Source: https://huggingface.co/papers/2607.04690

Abstract

WeintroducePAST-TIDE,ourstancedetectionsystemaddressingbothsubtasksoftheStanceNakbaSharedTaskatNakbaNLP@LREC-COLING2026.Themainideaisstatementtuning.Weredefinestanceascloze-stylemaskedlanguagemodeling(MLM),lettingaverbalizermaplabelwordstostancecategoriesthroughthepre-trainedMLMheadratherthanappendingarandomlyinitializedclassificationhead.Wecomplementthiswithprototypicalcontrastivelearning,whichuseslearnableclassprototypesforbatch-sizeindependentcontrastivetraining,andtopic-conditionallayernormalizationforcross-topicArabicstancedetection.PAST-TIDEachievesmacro-F1scoresof0.75forSubtaskAand0.74forSubtaskBontheofficialleaderboard,indicatingthatminimalarchitecturaladditionstoapre-trainedmodelcanremaincompetitiveinlow-resourcesettings.

View arXiv pageView PDFGitHub1Add to collection

Get this paper in your agent:

hf papers read 2607\.04690

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/2607.04690 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.04690 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.04690 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