Xinxi Lyu

Hello!

I am a first-year PhD student at UIUC. I am a formal Predoctoral Young Investigator at Allen Institute for AI mentored by Pradeep Dasigi. I received my B.S. and M.S. in Computer Science at Paul G. Allen School of Computer Science & Engineering at University of Washington-Seattle, where I was mentored by Sewon Min and Hannaneh Hajishirzi.

My current research focus on learning, characterizing, and improving Large Language Models in stages of data, pretraining, and inference.

My Chinese name is written as 吕欣禧, which embodys the meaning of happiness.

selected publications

Frustratingly Simple Retrieval Improves Challenging, Reasoning-Intensive Benchmarks

Xinxi Lyu, Michael Duan, Rulin Shao, and 2 more authors

2025

Bib

@article{lyu2025compactds,
  title = {Frustratingly Simple Retrieval Improves Challenging, Reasoning-Intensive Benchmarks},
  author = {Lyu, Xinxi and Duan, Michael and Shao, Rulin and Koh, Pang and Min, Sewon},
  year = {2025},
}

HREF: Human Response-Guided Evaluation of Instruction Following in Language Models

Xinxi Lyu, Yizhong Wang, Hannaneh Hajishirzi, and 1 more author

2024

Bib

@article{lyu2024href,
  title = { HREF: Human Response-Guided Evaluation of Instruction Following in Language Models },
  author = {Lyu, Xinxi and Wang, Yizhong and Hajishirzi, Hannaneh and Dasigi, Pradeep},
  year = { 2024 },
}

Leveraging Set Assumption for Membership Inference in Language Models

Xinxi Lyu, Sewon Min, Ari Holtzman, and 4 more authors

In , 2024

Bib

@inproceedings{lyu2023set,
  title = { Leveraging Set Assumption for Membership Inference in Language Models },
  author = {Lyu, Xinxi and Min, Sewon and Holtzman, Ari and Mireshghallah, Niloofar and Elazar, Yanai and Hajishirzi, Hannaneh and Dasigi, Pradeep},
  year = { 2024 },
}

TULU 3: Pushing Frontiers in Open Language Model Post-Training

Nathan Lambert, Jacob Morrison, Valentina Pyatkin, and 8 more authors

2024

arXiv Bib

@article{lambert2024t,
  title = {TULU 3: Pushing Frontiers in Open Language Model Post-Training},
  author = {Lambert, Nathan and Morrison, Jacob and Pyatkin, Valentina and Huang, Shengyi and Ivison, Hamish and Brahman, Faeze and Miranda, Lester James V and Liu, Alisa and Dziri, Nouha and Lyu, Xinxi and others},
  booktitle = {arXiv preprint arXiv:2411.15124},
  year = {2024},
}

FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation

Sewon Min, Kalpesh Krishna, Xinxi Lyu, and 6 more authors

Empirical Methods in Natural Language Processing , 2023

arXiv Bib

@article{min2023factscore,
  title = {FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation},
  author = {Min, Sewon and Krishna, Kalpesh and Lyu, Xinxi and Lewis, Mike and Yih, Wen-tau and Koh, Pang Wei and Iyyer, Mohit and Zettlemoyer, Luke and Hajishirzi, Hannaneh},
  journal = {Empirical Methods in Natural Language Processing },
  year = {2023},
}

Z-ICL: Zero-Shot In-Context Learning with Pseudo-Demonstrations

Xinxi Lyu, Sewon Min, Iz Beltagy, and 2 more authors

the Association for Computational Linguistics , 2022

arXiv Bib

@article{lyu2022z,
  title = { Z-ICL: Zero-Shot In-Context Learning with Pseudo-Demonstrations },
  author = {Lyu, Xinxi and Min, Sewon and Beltagy, Iz and Zettlemoyer, Luke and Hajishirzi, Hannaneh},
  journal = { the Association for Computational Linguistics },
  year = {2022},
}

Rethinking the Role of Demonstrations: What makes In-context Learning Work?

Sewon Min, Xinxi Lyu, Ari Holtzman, and 4 more authors

Empirical Methods in Natural Language Processing , 2022

arXiv Bib

@article{min2022rethinking,
  title = { Rethinking the Role of Demonstrations: What makes In-context Learning Work? },
  author = {Min, Sewon and Lyu, Xinxi and Holtzman, Ari and Artetxe, Mikel and Lewis, Mike and Hajishirzi, Hannaneh and Zettlemoyer, Luke},
  journal = { Empirical Methods in Natural Language Processing },
  year = {2022},
}

Prompt Waywardness: The Curious Case of Discretized Interpretation of Continuous Prompts

Daniel Khashabi, Shane Lyu, Sewon Min, and 8 more authors

Conference of the North American Chapter of the Association for Computational Linguistics , 2021

arXiv Bib

@article{khashabi2021prompt,
  title = { {P}rompt {W}aywardness: The Curious Case of Discretized Interpretation of Continuous Prompts },
  author = {Khashabi, Daniel and Lyu, Shane and Min, Sewon and Qin, Lianhui and Richardson, Kyle and Singh, Sameer and Welleck, Sean and Hajishirzi, Hannaneh and Khot, Tushar and Sabharwal, Ashish and Choi, Yejin},
  journal = { Conference of the North American Chapter of the Association for Computational Linguistics },
  year = {2021},
}

selected publications

Contact