Saturated Transformers are Constant-Depth Threshold Circuits William Merrill author Ashish Sabharwal author Noah A Smith author 2022 text journal article Transactions of the Association for Computational Linguistics continuing MIT Press Cambridge, MA periodical academic journal merrill-etal-2022-saturated 10.1162/tacl_a_00493 https://aclanthology.org/2022.tacl-1.49/ 2022 10 843 856