Fast and Robust Early-Exiting Framework for Autoregressive Language Models with Synchronized Parallel Decoding

Fast and Robust Early-Exiting Framework for Autoregressive Language Models with Synchronized Parallel Decoding Sangmin Bae author Jongwoo Ko author Hwanjun Song author Se-Young Yun author 2023-12 text Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing Houda Bouamor editor Juan Pino editor Kalika Bali editor Association for Computational Linguistics Singapore conference publication bae-etal-2023-fast 10.18653/v1/2023.emnlp-main.362 https://aclanthology.org/2023.emnlp-main.362/ 2023-12 5910 5924