Abstract
Neural sequence models can generate highly fluent sentences but recent studies have also shown that they are also prone to hallucinate additional content not supported by the input, which can cause a lack of trust in the model. To better assess the faithfulness of the machine outputs, we propose a new task to predict whether each token in the output sequence is hallucinated conditioned on the source input, and collect new manually annotated evaluation sets for this task. We also introduce a novel method for learning to model hallucination detection, based on pretrained language models fine tuned on synthetic data that includes automatically inserted hallucinations. Experiments on machine translation and abstract text summarization demonstrate the effectiveness of our proposed approach -- we obtain an average F1 of around 0.6 across all the benchmark datasets and achieve significant improvements in sentence-level hallucination scoring compared to baseline methods. We also release our annotated data and code for future research at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2011.02593
