View on GitHub


Repository to track the progress in Natural Language Processing (NLP), including the datasets and the current state-of-the-art for the most common NLP tasks.

Stance detection

Stance detection is the extraction of a subject’s reaction to a claim made by a primary actor. It is a core part of a set of approaches to fake news assessment.



The RumourEval 2017 dataset has been used for stance detection in English (subtask A). It features multiple stories and thousands of reply:response pairs, with train, test and evaluation splits each containing a distinct set of over-arching narratives.

This dataset subsumes the large PHEME collection of rumors and stance, which includes German.

Model Accuracy Paper / Source
Kochkina et al. 2017 0.784 Turing at SemEval-2017 Task 8: Sequential Approach to Rumour Stance Classification with Branch-LSTM
Bahuleyan and Vechtomova 2017 0.780 UWaterloo at SemEval-2017 Task 8: Detecting Stance towards Rumours with Topic Independent Features

Go back to the README