method · first seen 6 Oct 2026
ultra-low-bit quantization
Also: quantized weights, 2-bit quantization, low-bit weight quantization, extreme quantization of large language models, low-bit quantization of large language models
Representing neural network weights with only a few bits per value to greatly reduce memory and compute cost.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Align, Then Correct: Training-Free Two-Stage Low-Rank Compensation for Extremely Quantized Large Language ModelsPapers · 6 Oct 2026 · Seobin Song, Geonho Lee, Janghwan Lee, Jungwook Choi
- Few Bits, One Law: Toward W2A4KV2Papers · 6 Oct 2026 · Kai Yi, Tarek Elgamal, Sruthikesh Surineni, Vignesh Vivekraja, Soumyadeep Ghosh,…
- SAGE: Pushing 2-Bit Vector-Quantized LLMs toward Real-World UsabilityPapers · 10 Oct 2026 · Xianglong Yan, Zenan Wu, Kewei Zhang, Kaicheng Yang, Yulun Zhang
- FidelityQ: Preserving Predictive Fidelity in Ultra-Low-Bit Language Model QuantizationPapers · 10 Oct 2026 · Yujie Chen, Weiyan Zheng, Yiwu Yao, Tianchi Hu, Bo Wang, Jiyu Guo, Wenzhan Li, J…
- Decode Late, Multiply Once: Rethinking the Dot Product for Quantized WeightsPapers · 22 Oct 2026 · Jonathan Tapson
- ThinkBit: Search Codewords, Recalibrate Scales, Select by ReasoningPapers · 26 Oct 2026 · Rhea Sanjay Sukthanker, Pashmina Cameron
Latest sources
- CoS-VQ: Coordinated Shared-Scale Vector Quantization for Large Language ModelsPapers · 27 Oct 2026 · Zukang Xu, Zhixiong Zhao, Xing Hu, Zhe Jiang, Xin Si, Dawei Yang
- ThinkBit: Search Codewords, Recalibrate Scales, Select by ReasoningPapers · 26 Oct 2026 · Rhea Sanjay Sukthanker, Pashmina Cameron
- Decode Late, Multiply Once: Rethinking the Dot Product for Quantized WeightsPapers · 22 Oct 2026 · Jonathan Tapson
- FidelityQ: Preserving Predictive Fidelity in Ultra-Low-Bit Language Model QuantizationPapers · 10 Oct 2026 · Yujie Chen, Weiyan Zheng, Yiwu Yao, Tianchi Hu, Bo Wang, Jiyu Guo, Wenzhan Li, J…
- SAGE: Pushing 2-Bit Vector-Quantized LLMs toward Real-World UsabilityPapers · 10 Oct 2026 · Xianglong Yan, Zenan Wu, Kewei Zhang, Kaicheng Yang, Yulun Zhang
- Few Bits, One Law: Toward W2A4KV2Papers · 6 Oct 2026 · Kai Yi, Tarek Elgamal, Sruthikesh Surineni, Vignesh Vivekraja, Soumyadeep Ghosh,…
- Align, Then Correct: Training-Free Two-Stage Low-Rank Compensation for Extremely Quantized Large Language ModelsPapers · 6 Oct 2026 · Seobin Song, Geonho Lee, Janghwan Lee, Jungwook Choi