Xin Xu xxupiano

👋 Hi, I'm Xin Xu, a CSE Ph.D. student at UCSD, advised by Prof. Julian McAuley.
👀 I'm interested in:
- Trustworthy NLP: LLM interpretability and and toxicity
- Knowledge Management: factuality and interpretability
- Music Science: music interpretation and controllable music generation
🎓 I obtained the CS master's degree from Zhejiang University, advised by Prof. Ningyu Zhang.
💻 I was fortunate to intern at Microsoft Research Asia, advised by Xu Tan.
🎵 I love music. I was a member of Wenqin Piano Society at Zhejiang University.
🐾 Web: https://xxupiano.github.io/
📬 How to reach me: xinxucs [at] ucsd [dot] edu

Research Work

Aging Benchmarks: When Benchmarks Age: Temporal Misalignment through Large Language Model Factuality Evaluation. [paper] LLM-eval @ NeurIPS 2025
BiasFreeBench: BiasFreeBench: a Benchmark for Mitigating Bias in Large Language Model Responses. [paper]
ReDis: Improving In-Context Learning with Reasoning Distillation. [paper]
BiasEdit: Debiasing Stereotyped Models via Model Editing. [paper] TrustNLP @ NAACL 2025
MachineSoM: Exploring Collaboration Mechanisms for LLM Agents: A Social Psychology View. [paper] ACL 2024

KnowEdit: A Comprehensive Study of Knowledge Editing for Large Language Models. [paper]
CKnowEdit: Benchmarking Chinese Knowledge Rectification in Large Language Models. [paper] ACL 2025
UnleashLLMRE: How to Unleash the Power of Large Language Models for Few-shot Relation Extraction? [paper] SustaiNLP 2023 @ ACL 2023
AdaKGC: Schema-adaptable Knowledge Graph Construction. [paper] EMNLP 2023 Findings
DeepKE: A Deep Learning Based Knowledge Extraction Toolkit for Knowledge Base Population. [paper] EMNLP 2022 System Demostrations
LREBench: Towards Realistic Low-resource Relation Extraction: A Benchmark with Empirical Baseline Study. [paper] EMNLP 2022 Findings