zohaib khan

resume | scholar | github | linkedin | email

I am a Masters student at the University of Michigan Ann-Arbor, advised by Prof. Lu Wang and Prof. Rada Mihalcea. My research focuses on on AI safety, LLM reasoning, and post-training methods. I am supported by the Achievement Fellowship at Michigan. Before this, I studied computer science at LUMS in Pakistan.

news

2026-08-20 Two papers accepted to EMNLP Main 2026: Countdown-Code and Camellia!
2026-07-06 Our red-teaming paper received the SAC Highlight Award!
2026-04-06 Our project on multilingual red-teaming has been accepted to ACL as an Oral!
2026-03-02 Countdown-Code has been accepted to ICLR 2026 SPOT!
2026-03-01 Our project on multilingual fact-checking is up on Nature!

selected publications

Full list on Google Scholar
(* denotes equal contribution)

To Lie or Not to Lie? Investigating The Selective Spread of Global Lies by LLMs
Zohaib Khan, Mustafa Dogan, ..., Michael J Ryan, Tarek Naous
[SAC Highlight Award; Oral] ACL Main 2026
arxiv | github
Countdown-Code: A Testbed for Studying The Emergence and Generalization of Reward Hacking
Muhammad Khalifa*, Zohaib Khan*, Omer Tafveez, Hao Peng, Lu Wang
ICLR SPOT 2026, EMNLP Main 2026
arxiv | github
Large language models show Dunning-Kruger-like effects in multilingual fact-checking
Ihsan Ayyub Qazi, Zohaib Khan, Abdullah Ghani, Agha Ali Raza, Zafar Ayyub Qazi, Wassay Sajjad, Ayesha Ali, Asher Javaid, Muhammad Abdullah Sohail, Abdul Hameed Azeemi
Nature Scientific Reports
nature
Plasticity vs. Rigidity: The Impact of Low-Rank Adapters on LLM Reasoning on a Micro-Budget
Zohaib Khan, Omer Tafveez, Zoha Hayat Bhatti
EACL 2026 SRW
arxiv | github