@article{octopusoptimizedkvcachefortransformersvi, title = {OCTOPUS: Optimized KV Cache for Transformers via Octahedral Parametrization Under optimal Squared error quantization}, author = {Mark Boss and Vikram Voleti and Simon Donné and Shimon Vainer}, year = {2026}, eprint = {2605.21226}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2605.21226}, }