LLM.int8(): 8-Bit Matrix Multiplication for Transformers at Scale (2022)(arxiv.org)125 points by tosh 3 years ago | 15 comments
Code-native generation of highly programmable 3D assets (2026)(arxiv.org)37 points by baigy 2 months ago | 35 comments
Language models are injective and hence invertible(arxiv.org)194 points by mazsa 1 year ago | 136 comments