@article{itq3shighfidelity3bitllminferenceviainte, title = {ITQ3_S: High-Fidelity 3-bit LLM Inference via Interleaved Ternary Quantization with Rotation-Domain Smoothing}, author = {Edward J. Yoon}, year = {2026}, eprint = {2603.27914}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2603.27914}, }