{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T07:32:38Z","timestamp":1774510358561,"version":"3.50.1"},"reference-count":56,"publisher":"IEEE","license":[{"start":{"date-parts":[[2026,1,31]],"date-time":"2026-01-31T00:00:00Z","timestamp":1769817600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,31]],"date-time":"2026-01-31T00:00:00Z","timestamp":1769817600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,1,31]]},"DOI":"10.1109\/hpca68181.2026.11408453","type":"proceedings-article","created":{"date-parts":[[2026,3,4]],"date-time":"2026-03-04T20:47:22Z","timestamp":1772657242000},"page":"1-15","source":"Crossref","is-referenced-by-count":0,"title":["GyRot: Leveraging Hidden Synergy Between Rotation and Fine-Grained Group Quantization for Low-Bit LLM Inference"],"prefix":"10.1109","author":[{"given":"Sangjin","family":"Kim","sequence":"first","affiliation":[{"name":"School of Electrical Engineering, Korea Advanced Institue of Science and Technology (KAIST)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yuseon","family":"Choi","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea Advanced Institue of Science and Technology (KAIST)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Byeongcheol","family":"Kim","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea Advanced Institue of Science and Technology (KAIST)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jungjun","family":"Oh","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea Advanced Institue of Science and Technology (KAIST)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hoi-Jun","family":"Yoo","sequence":"additional","affiliation":[{"name":"School of Electrical Engineering, Korea Advanced Institue of Science and Technology (KAIST)"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.52202\/079017-3180"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i05.6239"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA61900.2025.00084"},{"key":"ref4","article-title":"Fp4 all the way: Fully quantized training of 11 ms","author":"Chmiel","year":"2025","journal-title":"arXiv preprint"},{"key":"ref5","article-title":"Boolq: Exploring the surprising difficulty of natural yes\/no questions","author":"Clark","year":"2019","journal-title":"arXiv preprint"},{"key":"ref6","article-title":"Think you have solved question answering? try arc, the ai2 reasoning challenge","author":"Clark","year":"2018","journal-title":"arXiv preprint"},{"key":"ref7","first-page":"873","article-title":"Vs-quant: Per-vector scaled quantization for accurate low-precision neural network inference","volume-title":"Proceedings of Machine Learning and Systems (MLSys)","volume":"3","author":"Dai","year":"2021"},{"key":"ref8","first-page":"10271","article-title":"Pushing the limits of narrow precision inferencing at cloud scale with microsoft floating point","volume":"33","author":"Darvish Rouhani","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref9","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018","journal-title":"arXiv preprint"},{"key":"ref10","article-title":"Training dnns with hybrid block floating point","volume":"31","author":"Drumond","year":"2018","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref11","article-title":"Gptq: Accurate post-training quantization for generative pre-trained transformers","author":"Frantar","year":"2022","journal-title":"arXiv preprint"},{"key":"ref12","volume-title":"The language model evaluation harness","author":"Gao","year":"2024"},{"key":"ref13","author":"Grattafiori","year":"2024","journal-title":"The llama 3 herd of models"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3579371.3589038"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO56248.2022.00095"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA61900.2025.00086"},{"key":"ref17","article-title":"Ostquant: Refining large language model quantization with orthogonal and scaling transformations for better distribution fitting","author":"Hu","year":"2025","journal-title":"arXiv preprint"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA61900.2025.00059"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/JETCAS.2025.3558300"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/ISSCC49661.2025.10904594"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA59077.2024.00080"},{"key":"ref22","article-title":"Amxfp4: Taming activation outliers with asymmetric microscaling floating-point for 4-bit 1lm inference","author":"Lee","year":"2024","journal-title":"arXiv preprint"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.52202\/079017-2786"},{"key":"ref24","first-page":"87","article-title":"Awq: Activation-aware weight quantization for on-device 11 m compression and acceleration","volume-title":"Proceedings of machine learning and systems","volume":"6","author":"Lin","year":"2024"},{"key":"ref25","article-title":"Qserve: W4a8kv4 quantization and system co-design for efficient llm serving","author":"Lin","year":"2024","journal-title":"arXiv preprint"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1907.11692"},{"key":"ref27","article-title":"Spinquant: Llm quantization with learned rotations","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Liu"},{"key":"ref28","year":"2025","journal-title":"Dram power calculator"},{"key":"ref29","year":"2025","journal-title":"Nvidia tensorrt-llm"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2018.00063"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1145\/3695053.3730989"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/3579371.3589351"},{"key":"ref33","article-title":"Microscaling data formats for deep learning","author":"Rouhani","year":"2023","journal-title":"arXiv preprint"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/3474381"},{"key":"ref35","article-title":"Resq: Mixed-precision quantization of large language models with low-rank residuals","author":"Saxena","year":"2024","journal-title":"arXiv preprint"},{"key":"ref36","article-title":"Omniquant: Omnidirectionally calibrated quantization for large language models","author":"Shao","year":"2023","journal-title":"arXiv preprint"},{"key":"ref37","article-title":"Glu variants improve transformer","author":"Shazeer","year":"2020","journal-title":"arXiv preprint"},{"key":"ref38","article-title":"Pointer sentinel mixture models","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Stephen","year":"2016"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2023.127063"},{"key":"ref40","article-title":"Flatquant: Flatness matters for llm quantization","author":"Sun","year":"2024","journal-title":"arXiv preprint"},{"key":"ref41","volume-title":"Design compiler: Timing, area, power, & test optimization","year":"2025"},{"key":"ref42","article-title":"Gemma: Open models based on gemini research and technology","author":"Team","year":"2024","journal-title":"arXiv preprint"},{"key":"ref43","article-title":"Llama: Open and efficient foundation language models","author":"Touvron","year":"2023","journal-title":"arXiv preprint"},{"key":"ref44","article-title":"Llama 2: Open foundation and fine-tuned chat models","author":"Touvron","year":"2023","journal-title":"arXiv preprint"},{"key":"ref45","article-title":"Quip\\#: Even better llm quantization with hadamard incoherence and lattice codebooks","author":"Tseng","year":"2024","journal-title":"arXiv preprint"},{"key":"ref46","article-title":"Dfrot: Achieving outlier-free and massive activation-free for rotated 11 ms with refined rotation","author":"Xiang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref47","author":"Xiang","journal-title":"Duarot: Dual rotation for advanced outlier mitigation in rotated 11 ms."},{"key":"ref48","first-page":"38087","article-title":"Smoothquant: Accurate and efficient post-training quantization for large language models","volume-title":"International conference on machine learning","author":"Xiao","year":"2023"},{"key":"ref49","article-title":"Zeroquant-v2: Exploring post-training quantization in 11 ms from comprehensive study to low rank compensation","author":"Yao","year":"2023","journal-title":"arXiv preprint"},{"key":"ref50","article-title":"Rptq: Reorder-based post-training quantization for large language models","author":"Yuan","year":"2023","journal-title":"arXiv preprint"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO50266.2020.00071"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1472"},{"key":"ref53","article-title":"Opt: Open pre-trained transformer language models","author":"Zhang","year":"2022","journal-title":"arXiv preprint"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.3"},{"key":"ref55","first-page":"196","article-title":"Atom: Low-bit quantization for efficient and accurate 11 m serving","volume-title":"Proceedings of Machine Learning and Systems","volume":"6","author":"Zhao","year":"2024"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2020"}],"event":{"name":"2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)","location":"Sydney, Australia","start":{"date-parts":[[2026,1,31]]},"end":{"date-parts":[[2026,2,4]]}},"container-title":["2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11408404\/11408433\/11408453.pdf?arnumber=11408453","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T05:33:24Z","timestamp":1774503204000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11408453\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,31]]},"references-count":56,"URL":"https:\/\/doi.org\/10.1109\/hpca68181.2026.11408453","relation":{},"subject":[],"published":{"date-parts":[[2026,1,31]]}}}