Hey there 👋
Hey there 👋
My name's Keenan and I'm broadly interested in (1) data-centric approaches for characterizing language model behaviors and (2) methods to govern and control such behaviors. I'm largely motivated by a desire to ensure that Artificial Intelligence is safe, aligned, and beneficial to society.
I recently finished a six-month research internship at the Max Planck Institute for Intelligent Systems in Tübingen, Germany, where I had the opportunity to work under the invaluable supervision of Zhijing Jin and Bernhard Schölkopf. Before that, I earned a master's in Computational Statistics and Machine Learning at UCL, where I was lucky to be advised by Yihong Chen and Pontus Stenetorp. I previously completed my bachelor's at Georgetown University in Science, Technology, and International Affairs, with a minor in Computer Science, where I was fortunate to work with Christopher Lawrence and Michael Rossetti. I'm also particularly grateful for Xiang Ding and Colin McCormick, who helped ignite my interest in machine learning back in 2021.
Outside of work, I enjoy olympic-style recurve archery 🏹, road cycling 🚴, hiking 🥾, traveling 🌐, and taking photos (lots of photos) 📷. I especially love seeking out adventures and novel experiences.
I'm applying to PhD programs in the US for a fall 2026 start!
[Nov 2025] I'll be joining Cohere Labs in January as a Research Scholar!
[Nov 2025] I presented our work on LLM moral reasoning at EMNLP 2025 in Suzhou, China 🇨🇳.
[Oct 2025] I gave an oral presentation on human rights evals for LLMs at the NLP4Democracy workshop at COLM 2025 in Montreal, Canada 🇨🇦.
[Sep 2025] I started a project with the Supervised Program for Alignment Research advised by Arush Tagade and Shi Feng.
[July 2025] I'm honored to have been awarded an NSF CISE Graduate Fellowship.
[May 2025] I've graduated from UCL with a MSc in CSML with Distinction!
Are Language Models Consequentialist or Deontological Moral Reasoners?
Keenan Samway*, Max Kleiman-Weiner*, David Guzman Piedrahita, Rada Mihalcea, Bernhard Schölkopf, Zhijing Jin
In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP 2025)
[proceedings] [arXiv] [poster]
When Do Language Models Endorse Limitations on Human Rights Principles?
Keenan Samway, Nicole Miu Takagi, Rada Mihalcea, Bernhard Schölkopf, Ilias Chalkidis, Daniel Hershcovich, Zhijing Jin
Under Review in ARR
Presented at the NLP4Democracy workshop at COLM 2025 (oral)
[poster]
Preserving Historical Truth: Detecting Historical Revisionism in Large Language Models
Francesco Ortu*, Joeun Yook*, Keenan Samway, Bernhard Schölkopf, Alberto Cazzaniga, Rada Mihalcea, and Zhijing Jin
Under Review in ARR
NLP for Social Good: A Survey and Outlook of Challenges, Opportunities and Responsible Deployment
Antonia Karamolegkou, Angana Borahᵅ, Eunjung Choᵅ, Sagnik Ray Choudhuryᵅ, Martina Gallettiᵅ, Pranav Guptaᵅ, Oana Ignatᵅ, Priyanka Karguptaᵅ, Neema Kotonyaᵅ, Hemank Lambaᵅ, Sun-Joo Leeᵅ, Arushi Manglaᵅ, Ishani Mondalᵅ, Fatima Zahra Moudakirᵅ, Deniz Nazarovaᵅ, Poli Nemkovaᵅ, Dina Pisarevskayaᵅ, Naquee Rizwanᵅ, Nazanin Sabriᵅ, Keenan Samwayᵅ, Dominik Stammbachᵅ, Anna Steinbergᵅ, David Tomásᵅ, Steven R Wilsonᵅ, Bowen Yiᵅ, Jessica H Zhuᵅ, Arkaitz Zubiagaᵅ, Anders Søgaard, Alexander Fraser, Zhijing Jin, Rada Mihalcea, Joel R. Tetreault, and Daryna Dementieva
Under Review in ARR
[arXiv]
NSF CISE Graduate Fellowship, CSGrad4US (2026)
UCL MSc in CSML with Distinction (2024)