Research

Recent publications

• Amplifying Trans and Nonbinary Voices: A Community-Centred Harm Taxonomy for LLMs. Eddie L. Ungless, Sunipa Dev, Cynthia L. Bennett, Rebecca Gulotta, Jasmijn Bastings, Remi Denton. ACL 2025.
• A decade of gender bias in machine translation. Beatrice Savoldi, Jasmijn Bastings, Luisa Bentivogli, Eva Vanmassenhove. Cell Patterns, June 2025. Cover article.
• MiTTenS: A Dataset for Evaluating Gender Mistranslation. Kevin Robinson, Sneha Kudugunta, Romina Stella, Sunipa Dev, Jasmijn Bastings. EMNLP 2024.
• Low-Rank Adaptation for Multilingual Summarization: An Empirical Study. Chenxi Whitehouse, Fantine Huot, Jasmijn Bastings, Mostafa Dehghani, Chu-Cheng Lin, Mirella Lapata. Findings of NAACL, 2024.
• Diagnosing AI explanation methods with folk concepts of behavior. Alon Jacovi, Jasmijn Bastings, Sebastian Gehrmann, Yoav Goldberg, Katja Filippova. JAIR 2023.
• Dissecting Recall of Factual Associations in Auto-Regressive Language Models. Mor Geva, Jasmijn Bastings, Katja Filippova, Amir Globerson. EMNLP 2023.

See my full publication list on my Google Scholar profile.

Highlighted publications

• We Need To Talk About Random Splits. Anders Søgaard, Sebastian Ebert, Jasmijn Bastings, Katja Filippova. 🏆 Honorable Mention for Best Short Paper. EACL, 2021.
• The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?. Jasmijn Bastings, Katja Filippova. BlackboxNLP 2020.
• Interpretable neural predictions with differentiable binary variables. Jasmijn Bastings, Wilker Aziz, Ivan Titov. ACL 2019.
• Joey NMT: A Minimalist NMT Toolkit for Novices. Julia Kreutzer, Jasmijn Bastings, Stefan Riezler. EMNLP 2019. [code]
• Graph convolutional encoders for syntax-aware neural machine translation Jasmijn Bastings, Ivan Titov, Wilker Aziz, Diego Marcheggiani, Khalil Sima’an. EMNLP 2017.

You can find a full list of my publications on my Google Scholar profile.

Code

• Interpretable Neural Predictions with Differentiable Binary Variables contains the HardKuma distribution that allows (hybrid) binary samples (with true zeros and ones) that allow gradients to pass through.
• Joey NMT is an easy-to-use, educational, and benchmarked NMT toolkit for novices that I developed with Julia Kreutzer and is currently maintained by Mayumi Ohta.
• FREVAL is an all-fragments parser evaluation metric that I developed with Khalil Sima’an.

Talks

• Workshop Measuring and mitigating bias in AI, Amsterdam, 2026. Building a Community-Centred Harm Taxonomy for LLMs.
• MilaNLP, Milan, 2024. Bits, Bats, and Bots: Deconstructing Gender in Language Technology.
• EMNLP, Blackbox NLP (Virtual), 2020. The Elephant in the Interpretability Room. (PDF)
• ACL, Florence, 2019. Interpretable Neural Predictions with Differentiable Binary Variables (Google Slides)
• EMNLP, Copenhagen, 2017. Graph Convolutional Encoders for Syntax-Aware Neural Machine Translation (Google Slides)