Is this you? As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.
Claim your profile
Articles by van Genabith
Computer Science > Computation and Language
Cultural evaluation coverage and robustness in language models are difficult to diagnose because pretraining corpora and cultural benchmarks are rarely indexed with comparable metadata. Benchmarks increasingly target culturally situated phenomena at the level of languages, regions, and locale-specific practices, while web-scale corpora are usually organized only by language.
Computer Science > Computation and Language
[Submitted on 8 Sep 2026] Title:Tracing Stereotypes from Representation to Output in Multilingual LLMs View a PDF of the paper titled Tracing Stereotypes from Representation to Output in Multilingual LLMs, by Ariun-Erdene Tumurchuluun and 4 other authors View PDF Abstract:Multilingual LLMs show stereotype-related behavior that varies across languages, but behavioral scores do not show where the relevant information is represented or how it affects the output.
Computer Science > Computation and Language
This study examines the compositionality of steering vectors for language and behavioral control in large language models. Focusing on language, jailbreak, and conciseness, we investigate whether additive, training-free composition of attribute steering vectors can preserve the intended steering effect of each attribute, across four instruction-tuned models from two model families and two size scales.
Computer Science > Computation and Language
Tokenization in language models is treated by default as an input preprocessing decision. We argue that this framing is incomplete: in autoregressive models, tokenizer granularity determines what the model must resolve in a single forward pass, and therefore the supervision signal it receives.
Computer Science > Computation and Language
Multilingual large language models (mLLMs) achieve strong performance in machine translation, yet our understanding of the mechanisms by which they transform representations from one language to another remains incomplete. Prior work suggests that translation decomposes into separable processes within an mLLM, where conceptual content is first represented independently, followed by a production into language-specific form.
Computer Science > Computation and Language
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Computer Science > Computation and Language
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Computer Science > Computation and Language
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Computer Science > Computation and Language
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Computer Science > Computation and Language
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
When Flores Bloomz Wrong: Cross-Direction Contamination in Machine Translation Evaluation
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
CLaS-Bench: A Cross-Lingual Alignment and Steering Benchmark
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
PragExTra: A Multilingual Corpus of Pragmatic Explicitation in Translation
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Seeing, Signing, and Saying: A Vision-Language Model-Assisted Pipeline for Sign Language Data Acquisition and Curation from Social Media
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
[2510.19413] Spatio-temporal Sign Language Representation and Translation
arXiv:2510.19413 (cs) View PDF HTML (experimental) Subjects: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2510.19413 [cs.CL] (or arXiv:2510.19413v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2510.19413 Journal reference: published at WMT 2022 Submission history From: Yasser Hamidullah Yasser H.
[2510.18439] Grounding or Guessing? Visual Signals for Detecting Hallucinations in Sign Language Translation
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
[2509.26415] Automatic Fact-checking in English and Telugu
arXiv:2509.26415 (cs) View PDF HTML (experimental) Subjects: Computation and Language (cs.CL) Cite as: arXiv:2509.26415 [cs.CL] (or arXiv:2509.26415v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2509.26415 Submission history From: Ravi Kiran Chikkala [ view email] [v1] Tue, 30 Sep 2025 15:39:34 UTC (5,463 KB) Bibliographic Tools Bibliographic Explorer Toggle Bibliographic Explorer () Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What...
When Scale Meets Diversity: Evaluating Language Models on Fine-Grained Multilingual Claim Verification
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
[2505.16956] On Multilingual Encoder Language Model Compression for Low-Resource Languages
Work on one of the world's most important websites and make an impact on open science.
AutoPsyC: Automatic Recognition of Psychodynamic Conflicts from Semi-structured Interviews with Large Language Models
[Submitted on 27 Mar 2025] Title:AutoPsyC: Automatic Recognition of Psychodynamic Conflicts from Semi-structured Interviews with Large Language Models View a PDF of the paper titled AutoPsyC: Automatic Recognition of Psychodynamic Conflicts from Semi-structured Interviews with Large Language Models, by Sayed Muddashir Hossain and 4 other authors View PDF HTML (experimental) Abstract:Psychodynamic conflicts are persistent, often unconscious themes that shape a person's behaviour and experiences.
[2409.14097] Probing Context Localization of Polysemous Words in Pre-trained Language Model Sub-Layers
[Submitted on 21 Sep 2024] Title:Probing Context Localization of Polysemous Words in Pre-trained Language Model Sub-Layers View a PDF of the paper titled Probing Context Localization of Polysemous Words in Pre-trained Language Model Sub-Layers, by Soniya Vijayakumar and 1 other authors View PDF HTML (experimental) Abstract:In the era of high performing Large Language Models, researchers have widely acknowledged that contextual word representations are one of the key drivers in achieving top...
LLMCheckup: Conversational Examination of Large Language Models via Interpretability Tools and Self-Explanations
[Submitted on 23 Jan 2024 ( v1 ), last revised 24 Apr 2024 (this version, v2)] Title:LLMCheckup: Conversational Examination of Large Language Models via Interpretability Tools and Self-Explanations View a PDF of the paper titled LLMCheckup: Conversational Examination of Large Language Models via Interpretability Tools and Self-Explanations, by Qianli Wang and 5 other authors View PDF HTML (experimental) Abstract:Interpretability tools that offer explanations in the form of a dialogue have...
[2312.06514] Where exactly does contextualization in a PLM happen?
arXiv:2312.06514 (cs) Download PDF HTML (experimental) Comments: Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI) Cite as: arXiv:2312.06514 [cs.CL] (or arXiv:2312.06514v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2312.06514 Submission history From: Soniya Vijayakumar [ view email] [v1] Mon, 11 Dec 2023 16:39:52 UTC (96 KB) Bibliographic Tools Bibliographic Explorer Toggle Bibliographic Explorer () Litmaps Toggle Litmaps (What is Litmaps?) scite.ai...
Measuring Spurious Correlation in Classification: 'Clever Hans' in Translationese
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
RescueSpeech: A German Corpus for Speech Recognition in Search and Rescue Domain
[Submitted on 6 Jun 2023 ( v1 ), last revised 25 Sep 2023 (this version, v3)] Title:RescueSpeech: A German Corpus for Speech Recognition in Search and Rescue Domain Download a PDF of the paper titled RescueSpeech: A German Corpus for Speech Recognition in Search and Rescue Domain, by Sangeet Sagar and 4 other authors Download PDF Abstract:Despite the recent advancements in speech recognition, there are still difficulties in accurately transcribing conversational and emotional speech in noisy...
[2304.14796] Are the Best Multilingual Document Embeddings simply Based on Sentence Embeddings?
arXiv:2304.14796 (cs) Download a PDF of the paper titled Are the Best Multilingual Document Embeddings simply Based on Sentence Embeddings?, by Sonal Sannigrahi and 2 other authors Download PDF Submission history From: Sonal Sannigrahi [ view email] [v1] Fri, 28 Apr 2023 12:11:21 UTC (3,446 KB) Bibliographic Tools Bibliographic Explorer Toggle Bibliographic Explorer () Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media...
[2304.10216] Exploring Paracrawl for Document-level Neural Machine Translation
arXiv:2304.10216 (cs) Download a PDF of the paper titled Exploring Paracrawl for Document-level Neural Machine Translation, by Yusser Al Ghussin and 2 other authors Download PDF Submission history From: Yusser Al Ghussin [ view email] [v1] Thu, 20 Apr 2023 11:21:34 UTC (6,621 KB) Bibliographic Tools Bibliographic Explorer Toggle Bibliographic Explorer () Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media DagsHub Toggle...
NAPG: Non-Autoregressive Program Generation for Hybrid Tabular-Textual Question Answering
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
[2007.06257] Rewiring the Transformer with Depth-Wise LSTMs
[Submitted on 13 Jul 2020 ( v1 ), last revised 4 Apr 2024 (this version, v2)] Title:Rewiring the Transformer with Depth-Wise LSTMs View a PDF of the paper titled Rewiring the Transformer with Depth-Wise LSTMs, by Hongfei Xu and Yang Song and Qiuhui Liu and Josef van Genabith and Deyi Xiong View PDF HTML (experimental) Abstract:Stacking non-linear layers allows deep neural networks to model complicated functions, and including residual connections in Transformer layers is beneficial for...
Show More
loading
Actions
Is this you?
As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.Get in touch with van
Contact van, search articles and posts on X, monitor coverage, and track replies from one place.
Learn more about Muck Rack