Prompting is now the primary way to utilize the multitask capabilities o...
Retrieval-augmented in-context learning has emerged as a powerful approa...
Many real-world applications of language models (LMs), such as code
auto...
Likelihood, although useful as a training loss, is a poor search objecti...
Controlling the behavior of language models (LMs) without re-training is...
Fine-tuning is the de facto way to leverage large pretrained language mo...
Narrative generation is an open-ended NLP task in which a model generate...
Text generation often requires high-precision output that obeys task-spe...
Pre-trained word embeddings like ELMo and BERT contain rich syntactic an...
Treebanks traditionally treat punctuation marks as ordinary words, but
l...