Old News

Our paper about quantization's effect on factual knowledge recall is finally accepted (Oral) at TrustNLP workshop co-located at ACL 2026!

🌟I am thrilled to shared that two papers are accepted at ACL 2026 about (1) multilingual counterfactual evaluation (Main); (2) actionable mechanistic interpretability survey (Findings)! See you in San Diego 🇺🇸

🤖 One co-authored paper in collaboration with Uni Marburg is in submission. We propose SimXAI, a lightweight tool-augmented LLM-based conversational layer for ConvXAI systems, which enables better interaction quality and human understanding.

Two first-authored and three co-authored papers in submission about (1) how quantization affects self-explanations❓collaborated with Pepa Atanasova; (2) feedback-driven iterative counterfactual refinement; (3) NLG evaluation trend survey led by Jing Yang; (4) claim-level grounding traceable generation led by Bohao Chu; (5) mechanistic interpretability application survey led by Hengyuan Zhang.

📝 One paper about multilingual counterfactual evaluation in collaboration with LMU and Uni Marburg in submission.

🌱 Start my research stay in the CopeNLU group led by Prof. Dr. Isabelle Augenstein in Denmark🧜. Thanks for hosting me and providing a nice starting point for our collaborations!

💥Delighted to share that (1) Multilingual Datasets for Custom Input Extraction and Explanation Requests Parsing in Conversational XAI Systems accepted at EMNLP 2025 Findings; (2) Truth or Twist? Optimal Model Selection for Reliable Label Flipping Evaluation in LLM-based Counterfactuals accepted at INLG 2025. See you in Hanoi, Vietnam 🇻🇳 and Suzhou, China 🇨🇳! ✈️

💫 Three first-authored papers in submission about (1) multilingual explanation request parsing and custom input extraction in ConvXAI; (2) counterfactual validity evaluation for counterfactual data augmentation; (3) how does quantization affect model explainability and interpretability.

The 4th first-authored paper titled FitCF: A Framework for Automatic Feature Importance-guided Counterfactual Example Generation accepted at ACL 2025 Findings. See you in Vienna, Austria 🇦🇹! 🍻

🚥 Thrilled to announce that I will be conducting a research stay in September at the CopeNLU group, led by Isabelle Augenstein, funded by the Danish Data Science Academy(DDSA).

I gave an invited talk at Efficient and Explainable NLP (E&E) Group, MLT Lab, DFKI about FitCF paper.

🎉 Exciting News from AI Founder Fellowship Program: I'm glad to announce that I have received funding granted by German Academy of Technical Sciences (acatech) and Federal Ministry for Digital and Transport (BMDV), supporting PhD students with up to 32 Mio. Euro over 9 months for their doctoral research projects/start-up ideas.

I will be at COLING 2025 to present our Cross-Refine paper. See you in Abu Dhabi!

🌴 One new paper in submission about feature importance-guided counterfactual example generation (FitCF). The preprint is now available on arXiv!

The 3rd first-authored paper titled Cross-Refine: Improving Natural Language Explanation Generation by Learning in Tandem accepted at COLING 2025, which will be hold in Abu Dhabi, United Arab Emirates🏜️!

I will be in Miami at EMNLP 2024 to present our CoXQL paper. See you in Miami!

One co-authored paper about LLM self-explanation alignment is in submission!

The 2nd first-authored paper CoXQL: A Dataset for Parsing Explanation Requests in Conversational XAI Systems accepted at EMNLP 2024 Findings, which will be hold in Miami, USA 🗽!

One new paper in submission: an NLG paper for improving explanations via cross-refinement (Cross-Refine).

One co-authored paper - Towards Modeling and Evaluating Instructional Explanations in Teacher-Student Dialogues accepted to ACM GoodIT 2024.

One new paper in submission: A question answering (resource) paper on the conversational parsing explanation requests (CoXQL).

The 1st first-authored paper LLMCheckup: Conversational Examination of Large Language Models via Interpretability Tools and Self-Explanations accepted to NAACL 2024 HCI+NLP workshop🛕!

Two new papers in submission: A follow-up to InterroLang on the conversational examination of self-explaining LLMs (LLMCheckup) and a resource and evaluation paper on instructional explanations in teacher-student dialogues (ReWIRED).

InterroLang: Exploring NLP Models and Datasets through Dialogue-based Explanations accepted to EMNLP 2023 Findings🦁!