People

I am always glad to hear from students interested in reasoning, information extraction, or AI for Science — see Contact.

PhD students I have worked with:

• Samuel Lewis-Lim — University of Sheffield
  Chain-of-thought faithfulness, monitorbility, and efficiency.
  Analysing Chain of Thought Dynamics (EMNLP 2025) · Can Confidence Estimates Decide When CoT Is Necessary? (Findings of AACL–IJCNLP 2026)

• Mingzi Cao — University of Sheffield
  Large Language Models merging, up-scaling, and editing.
  Fundamental Reasoning Paradigms Induce Out-of-Domain Generalization (Findings of ACL 2026)

• Constantinos Karouzos — University of Sheffield
  Preference tuning, generalisation and diversity under domain shift.
  An Empirical Study on Preference Tuning Generalization and Diversity Under Domain Shift (EMNLP 2026)

• Vynska Amalia Permadi — University of Sheffield
  Multi-hop question answering and cultural understanding.
  No Shortcuts to Culture (TACL 2026)

• Danqin Zhao — University of Warwick
  Relation extraction and graph construction in Psychology literature.
  EmpiriGraph-Psy: A Dataset and LLM Pipeline for Extracting Empirical Relation Graphs from Psychology Abstracts

• Chen Lv — University of Warwick
  Conversation generation and AI for law enforcement.
  ConVAWG: A Retrieval-Grounded Framework for Controlled Synthetic Dialogue Generation in Violence Against Women and Girls

• Ahmed Alajrami — University of Sheffield
  Instruction tuning and robustness to noisy supervision.
  Fine-Tuning on Noisy Instructions (AACL–IJCNLP 2025)