{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,25]],"date-time":"2026-02-25T21:57:30Z","timestamp":1772056650008,"version":"3.50.1"},"reference-count":54,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"1","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100003725","name":"National Research Foundation of Korea","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100003725","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Korea government","award":["2022R1A3B1077720"],"award-info":[{"award-number":["2022R1A3B1077720"]}]},{"name":"BK21 FOUR program of the Education and Research Program for Future ICT Pioneers"},{"name":"Seoul National University in 2025, Institute of Information"},{"name":"Korea government","award":["RS-2021-II211343"],"award-info":[{"award-number":["RS-2021-II211343"]}]},{"name":"Artificial Intelligence Graduate School Program"},{"DOI":"10.13039\/501100002551","name":"Seoul National University","doi-asserted-by":"publisher","award":["RS-2022-II220959"],"award-info":[{"award-number":["RS-2022-II220959"]}],"id":[{"id":"10.13039\/501100002551","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Institute of Information & communications Technology Planning"},{"name":"Leading Generative AI Human Resources Development","award":["IITP-2025-RS-2024-00397085"],"award-info":[{"award-number":["IITP-2025-RS-2024-00397085"]}]},{"name":"Korea Government"},{"name":"National Information Society Agency","award":["2025-AI Data-wi43"],"award-info":[{"award-number":["2025-AI Data-wi43"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE J. Sel. Top. Signal Process."],"published-print":{"date-parts":[[2026,1]]},"DOI":"10.1109\/jstsp.2026.3657641","type":"journal-article","created":{"date-parts":[[2026,1,28]],"date-time":"2026-01-28T20:59:34Z","timestamp":1769633974000},"page":"90-98","source":"Crossref","is-referenced-by-count":0,"title":["Beyond Language-Specific Neurons: The Challenge of Identifying Speech-Specific Neurons in Multimodal LLMs"],"prefix":"10.1109","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0009-0007-8142-7768","authenticated-orcid":false,"given":"Nohil","family":"Park","sequence":"first","affiliation":[{"name":"Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-3419-1895","authenticated-orcid":false,"given":"Che Hyun","family":"Lee","sequence":"additional","affiliation":[{"name":"Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-0424-7101","authenticated-orcid":false,"given":"Jiheum","family":"Yeom","sequence":"additional","affiliation":[{"name":"Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-1402-2425","authenticated-orcid":false,"given":"Heeseung","family":"Kim","sequence":"additional","affiliation":[{"name":"Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2367-197X","authenticated-orcid":false,"given":"Sungroh","family":"Yoon","sequence":"additional","affiliation":[{"name":"Department of Electrical and Computer Engineering, Interdisciplinary Program in Artificial Intelligence, ASRI, INMC, ISRC, and SOFT Foundry, Seoul National University, Seoul, South Korea"}]}],"member":"263","reference":[{"key":"ref1","first-page":"1877","article-title":"Language models are few-shot learners","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Brown","year":"2020"},{"key":"ref2","article-title":"LLaMA: Open and efficient foundation language models","author":"Touvron","year":"2023"},{"key":"ref3","article-title":"Mistral 7b","author":"Jiang","year":"2023"},{"key":"ref4","article-title":"Qwen2 technical report","author":"Yang","year":"2024"},{"key":"ref5","article-title":"The Llama 3 herd of models","author":"Grattafiori","year":"2024"},{"key":"ref6","article-title":"GPT-4 technical report","author":"Achiam","year":"2024"},{"key":"ref7","doi-asserted-by":"crossref","first-page":"8440","DOI":"10.18653\/v1\/2020.acl-main.747","article-title":"Unsupervised cross-lingual representation learning at scale","volume-title":"Proc. 58th Annu. Meet. Assoc. Comput. Linguistics","author":"Conneau","year":"2020"},{"key":"ref8","first-page":"483","article-title":"mT5: A massively multilingual pre-trained text-to-text transformer","volume-title":"Proc. Conf. North Amer. Chapter Assoc. Comput. Linguistics-Human Lang. Technol.","author":"Xue","year":"2021"},{"key":"ref9","first-page":"9019","article-title":"Few-shot learning with multilingual generative language models","volume-title":"Proc. Conf. Empirical Methods Natural Lang. Process.","author":"Lin","year":"2022"},{"key":"ref10","doi-asserted-by":"crossref","first-page":"15991","DOI":"10.18653\/v1\/2023.acl-long.891","article-title":"Crosslingual generalization through multitask finetuning","volume-title":"Proc. 61st Annu. Meet. Assoc. Comput. Linguistics (Volume 1: Long Papers)","author":"Muennighoff","year":"2023"},{"key":"ref11","first-page":"34892","article-title":"Visual instruction tuning","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"36","author":"Liu","year":"2023"},{"key":"ref12","article-title":"Qwen-VL: A versatile vision-language model for understanding, localization, text reading, and beyond","author":"Bai","year":"2024"},{"key":"ref13","article-title":"Qwen2.5-omni technical report","author":"Xu","year":"2025"},{"key":"ref14","first-page":"5701","article-title":"Language-specific neurons: The key to multilingual capabilities in large language models","volume-title":"Proc. 62nd Annu. Meeting Assoc. Comput. Linguistics","author":"Tang","year":"2024"},{"key":"ref15","first-page":"6801","article-title":"MMNeuron: Discovering neuron-level domain-specific interpretation in multimodal large language model","volume-title":"Proc. Conf. Empirical Methods Natural Lang. Process.","author":"Huo","year":"2024"},{"key":"ref16","first-page":"6919","article-title":"On the multilingual ability of decoder-based pre-trained language models: Finding and controlling language-specific neurons","volume-title":"Proc. Conf. North Amer. Chapter Assoc. Comput. Linguistics-Human Lang. Technol.","author":"Kojima","year":"2024"},{"key":"ref17","doi-asserted-by":"crossref","first-page":"15366","DOI":"10.18653\/v1\/2024.acl-long.820","article-title":"Do Llamas work in English? On the latent language of multilingual transformers","volume-title":"Proc. 62nd Annu. Meet. Assoc. Comput. Linguistics (Volume 1: Long Papers)","author":"Wendler","year":"2024"},{"key":"ref18","first-page":"2289","article-title":"From redundancy to relevance: Enhancing explainability in multimodal large language models","volume-title":"Proc. Annu. Conf. Nations Americas Chapter Assoc. Comput. Linguistics","author":"Zhang","year":"2025"},{"key":"ref19","first-page":"122867","article-title":"Towards neuron attributions in multi-modal large language models","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Fang","year":"2024"},{"key":"ref20","first-page":"15296","article-title":"How do large language models handle multilingualism?","volume-title":"Proc. 38th Annu. Conf. Neural Inf. Process. Syst.","author":"Zhao","year":"2024"},{"key":"ref21","article-title":"GPT-4o system card","author":"Hurst","year":"2024"},{"key":"ref22","article-title":"AudioPaLM: A large language model that can speak and listen","author":"Rubenstein","year":"2023"},{"key":"ref23","first-page":"18617","article-title":"LLaMA-omni 2: LLM-based real-time spoken chatbot with autoregressive streaming speech synthesis","volume-title":"Proc. 63rd Annu. Meeting Assoc. Comput. Linguistics","author":"Fang","year":"2025"},{"key":"ref24","article-title":"Kimi-audio technical report","author":"Ding","year":"2025"},{"key":"ref25","article-title":"Qwen2.5-math technical report","author":"Yang","year":"2025"},{"key":"ref26","first-page":"5484","article-title":"Transformer feed-forward layers are key-value memories","volume-title":"Proc. Conf. Empirical Methods Natural Lang. Process.","author":"Geva","year":"2021"},{"key":"ref27","first-page":"6000","article-title":"Attention is all you need","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Vaswani","year":"2017"},{"key":"ref28","doi-asserted-by":"crossref","first-page":"331","DOI":"10.18653\/v1\/P19-1032","article-title":"Adaptive attention span in transformers","volume-title":"Proc. 57th Annu. Meeting Assoc. Comput. Linguistics","author":"Sukhbaatar","year":"2019"},{"key":"ref29","first-page":"8493","article-title":"Knowledge neurons in pretrained transformers","volume-title":"Proc. 60th Annu. Meeting Assoc. Comput. Linguistics","author":"Dai","year":"2022"},{"key":"ref30","first-page":"17359","article-title":"Locating and editing factual associations in GPT","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Meng","year":"2022"},{"key":"ref31","first-page":"256","article-title":"Learnable privacy neurons localization in language models","volume-title":"Proc. 62nd Annu. Meeting Assoc. Comput. Linguistics","author":"Chen","year":"2024"},{"key":"ref32","first-page":"1288","article-title":"Neurons in large language models: Dead, n-gram, positional","volume-title":"Proc. Assoc. Comput. Linguistics","author":"Voita","year":"2024"},{"key":"ref33","article-title":"Universal neurons in GPT-2 language models","author":"Gurnee","year":"2024","journal-title":"IEEE Trans. Mach. Learn. Res."},{"key":"ref34","article-title":"BLOOM: A 176B-parameter open-access multilingual language model","author":"Workshop","year":"2022"},{"key":"ref35","article-title":"Miner: Mining the underlying pattern of modality-specific neurons in multimodal large language models","author":"Huang","year":"2024"},{"key":"ref36","article-title":"Qwen technical report","author":"Bai","year":"2023"},{"key":"ref37","first-page":"15757","article-title":"SpeechGPT: Empowering large language models with intrinsic cross-modal conversational abilities","volume-title":"Proc. Assoc. Comput. Linguistics: Empir. Methods Natural Lang. Process.","author":"Zhang","year":"2023"},{"key":"ref38","first-page":"131072","article-title":"Paralinguistics-aware speech-empowered large language models for natural conversation","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Kim","year":"2024"},{"key":"ref39","first-page":"30","article-title":"SpiRit-LM: Interleaved spoken and written language model","volume-title":"Trans. Assoc. Comput. Linguistics","volume":"13","author":"Nguyen","year":"2025"},{"key":"ref40","article-title":"Qwen2-audio technical report","author":"Chu","year":"2024"},{"key":"ref41","article-title":"LLaMA-Omni: Seamless speech interaction with large language models","volume-title":"Proc. 13th Int. Conf. Learn. Representations","author":"Fang","year":"2025"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i22.34557"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/TASLPRO.2025.3533359"},{"key":"ref44","article-title":"Scaling speech-text pre-training with synthetic interleaved data","volume-title":"Proc. 13th Int. Conf. Learn. Representations","author":"Zeng","year":"2025"},{"key":"ref45","article-title":"GLM-4-voice: Towards intelligent and human-like end-to-end spoken chatbot","author":"Zeng","year":"2024"},{"key":"ref46","article-title":"Phi-4-mini technical report: Compact yet powerful multimodal language models via mixture-of-LoRAs","author":"Abouelenin","year":"2025"},{"key":"ref47","article-title":"Baichuan-omni-1.5 technical report","author":"Li","year":"2025"},{"key":"ref48","first-page":"933","article-title":"Language modeling with gated convolutional networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Dauphin","year":"2017"},{"key":"ref49","article-title":"Spoken question answering and speech continuation using spectrogram-powered LLM","volume-title":"Proc. 12th Int. Conf. Learn. Representations","author":"Nachmani","year":"2024"},{"key":"ref50","first-page":"1601","article-title":"TriviaQA: A large scale distantly supervised challenge dataset for reading comprehension","volume-title":"Proc. 55th Annu. Meeting Assoc. Comput. Linguistics","author":"Joshi","year":"2017"},{"key":"ref51","article-title":"VoiceBench: Benchmarking LLM-based voice assistants","author":"Chen","year":"2024"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2015.7178964"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP43922.2022.9746828"},{"key":"ref54","first-page":"5775","article-title":"Multilingual knowledge editing with language-agnostic factual neurons","volume-title":"Proc. 31st Int. Conf. Comput. Linguistics","author":"Zhang","year":"2025"}],"container-title":["IEEE Journal of Selected Topics in Signal Processing"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/4200690\/11409413\/11363444.pdf?arnumber=11363444","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,25]],"date-time":"2026-02-25T20:56:37Z","timestamp":1772052997000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11363444\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1]]},"references-count":54,"journal-issue":{"issue":"1"},"URL":"https:\/\/doi.org\/10.1109\/jstsp.2026.3657641","relation":{},"ISSN":["1932-4553","1941-0484"],"issn-type":[{"value":"1932-4553","type":"print"},{"value":"1941-0484","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,1]]}}}