Influence Patterns for Explaining Information Flow in BERT [NeurIPS 2021]
Lu, Kaiji, Zifan Wang, Piotr Mardziel, and Anupam Datta. "Influence Patterns for Explaining Information Flow in BERT." NeurIPS, 2021.
Lu, Kaiji, Zifan Wang, Piotr Mardziel, and Anupam Datta. "Influence Patterns for Explaining Information Flow in BERT." NeurIPS, 2021.
Lu, Kaiji, Piotr Mardziel, Klas Leino, Matt Fredrikson, and Anupam Datta. "Influence Paths for Characterizing Subject-VerbNumber Agreement in LSTM Language Models.." ACL 2020