Zobrazeno 1 - 4
of 4
pro vyhledávání: '"Vijayakumar, Soniya"'
In the era of high performing Large Language Models, researchers have widely acknowledged that contextual word representations are one of the key drivers in achieving top performances in downstream tasks. In this work, we investigate the degree of co
Externí odkaz:
http://arxiv.org/abs/2409.14097
Pre-trained Language Models (PLMs) have shown to be consistently successful in a plethora of NLP tasks due to their ability to learn contextualized representations of words (Ethayarajh, 2019). BERT (Devlin et al., 2018), ELMo (Peters et al., 2018) an
Externí odkaz:
http://arxiv.org/abs/2312.06514
Pretrained language models (PLMs) form the basis of most state-of-the-art NLP technologies. Nevertheless, they are essentially black boxes: Humans do not have a clear understanding of what knowledge is encoded in different parts of the models, especi
Externí odkaz:
http://arxiv.org/abs/2311.08240
Autor:
Vijayakumar, Soniya
Publikováno v:
CEUR WS - CIKM 2022 Workshops Proceedings, Vol-3318, 2022, paper6
The field of natural language processing has reached breakthroughs with the advent of transformers. They have remained state-of-the-art since then, and there also has been much research in analyzing, interpreting, and evaluating the attention layers
Externí odkaz:
http://arxiv.org/abs/2302.09304