Gerrit Quaremba

I am a second-year PhD student in Computer Science at King’s College London and Imperial College London, aligned with the Safe & Trusted AI CDT and PHAWM. I am advised by Elena Simperl, Elizabeth Black (KCL), and Denny Vrandečić (Wikimedia).

Broadly, my research focuses on how AI can improve information verifiability, using ML and NLP with a focus on LLM-generated text detection and fact-checking. I'm especially interested in applications on Wikipedia, for example, to ensure that claims in need of an authoritative citation are flagged for editors.

I have published at top-tier venues such as ICLR, ACL, and EMNLP, and am currently working on zero-shot LLM-generated text detection and fact-verdict prediction for Wikipedia claims. If you are interested in collaborating, please reach out!

Previously, I obtained my master’s degree from the London School of Economics and worked as a research fellow at Harvard, Durham University, and the University of Cologne.

In my free time, I enjoy playing tennis and watching football, Formula 1, and tennis, especially my fellow Italian players.

first_name.surname [at] kcl.ac.uk  /  CV  /  Scholar  /  Github

News

• August 2026: Our paper Linear Probing Provides Robust and Efficient Detection of Machine-Generated Text has been accepted to EMNLP 2026 (main)!
• July 2026: I presented our paper on multilingual and cross-lingual citation-needed detection (see below) as an oral presentation at ACL 2026 in San Diego. The paper also received a SAC Highlight, an award given by the Senior Area Chairs to recognize one paper in each batch that they consider particularly outstanding.
• May 2026: Our paper, co-authored with Bogdan Grecu, Argchestrators at UZH Shared Task 2026: Efficient Argument Mining in UN Resolutions: A Sub-8B Pipeline using Agentic Debate and Heuristic Retrieval , won third place at the UZH Shared Task @ ArgMining Workshop 2026!
• April 2026: Our paper Multilingual and Cross-Lingual Citation Needed Detection on Wikipedia for Lower-Resource Languages has been accepted to ACL 2026 (main)!
• March 2026: We presented our work Multilingual and Cross-Lingual Citation Needed Detection on Wikipedia for Lower-Resource Languages at the the 13th Wiki Workshop 2026!!
• February 2026: An extended abstract of our paper Multilingual and Cross-Lingual Citation Needed Detection on Wikipedia for Lower-Resource Languages got accepted to the 13th Wiki Workshop 2026 (non-archival)!
• January 2026: Our paper TSM-Bench: Detecting LLM-Generated Text in Real-World Wikipedia Editing Practices has been accepted to ICLR 2026!
• September 2025: A book chapter I co-authored with Moritz Osnabruegge and Xiao Lu on Analyzing the Language of Legislation Using Natural Language Processing has been published in the book Language for Legislation and Legislation through Language by Taylor & Francis.
• July 2025: I presented our paper WETBench at ACL 2025 in Vienna, Austria!
• April 2025: Our paper WETBench: A Benchmark for Detecting Task-Specific Machine-Generated Text on Wikipedia has been accepted to the WikiNLP Workshop at ACL 2025!
• October 2024: I joined King’s College London and Imperial College London as a PhD student.

Site template cloned from Jon Barron.