24 datasets found

Text generation often requires high-precision output that obeys task-specific rules. This fine-grained control is difficult to enforce with off-the-shelf deep learning models.

The SHROOM dataset is a collection of data points containing task, input text, target text, and generated text. The dataset is used for hallucination detection in natural...

Radiology report generation aims to automatically generate a clinically accurate and coherent paragraph from the X-ray image, which could relieve radiologists from the heavy...

The dataset used in the paper for validation studies of automatic metrics in natural language generation evaluation

The Famous Keyword Twitter Replies dataset is a comprehensive collection of Twitter data that focuses on popular keywords and their associated replies.

The dataset used for the text summarization task, where a summarizer produces an utterance made up of one or multiple sentences to succinctly report the main content of a text.

The dataset used for the reference games task, where participants produce descriptions that allow comprehenders to identify the correct referent out of a set of candidates.

The dataset used in the paper is LLaMA, a large language model.

1-billion-word dataset

CMU-SE dataset

Chinese poetry generation dataset

The dataset used in the paper to evaluate the effectiveness of referenceless metrics for image accessibility.

The dataset used in the paper is ACE 2005, WebNLG, CoNLL, NYT, and FB15k-237. The ACE 2005 dataset is a collection of news articles, while WebNLG is a corpus used for natural...

Radiology report generation using cross-modal memory networks

VideoIC dataset for automatic live video commenting

Livebot dataset for automatic live video commenting

Sentiment-oriented Transformer-based Variational Autoencoder (So-TVAE) for Live Video Commenting

A hierarchical table dataset for question answering and natural language generation.

Controlling the behavior of language models (LMs) without re-training is a major open problem in natural language generation. We develop a new non-autoregressive language model...

The dataset used in the paper is not explicitly described, but it is mentioned that the authors used six datasets from three different NLG tasks.