Hello world! My name is Krisna, a PhD student in Natural Language Processing at MBZUAI. My research encompasses interpretability of language and multimodal language models (with main focus on training dynamic or learning in general), and explainable AI (XAI). Specifically, my research is to interpret the blackbox language models over the course of training to enable better approach for representation learning and also designing performant explainable and controllable language models to enable such models to be “debuggable” easily by researchers just like a computer programme. You can reach out to me through mahardika.ihsani@mbzuai.ac.ae or krisnaihsani@gmail.com.

News

• 2025.08: 🎉 My work on understanding the cross-lingual factual recall has been accepted in EMNLP’25 Findings. See you in Suzhou!
• 2025.08: 🎉 Starting my PhD in MBZUAI

• 2025.07: 🎉 1 first-authored paper on interpretability [1] and 2 co-authored papers [1][2] have been posted on arXiv!

• 2025.03: 🎉 1 co-authored paper on crowdsourced multimodal dataset [1] has been published on arXiv and also got accepted in ACL 2025