About

Excited to be starting my Student Researcher internship at Google MTV! Looking forward to meeting new friends and reconnecting with old ones!

I am a first -> second -year CS PhD student at University of Virginia, advised by Prof. Yu Meng. I am fortunate to be supported by UVA Provost's Fellowship and UVA Computer Science Scholar Fellowship.

Previously, I was a research assistant at MiuLab, National Taiwan University, advised by Prof. Yun-Nung (Vivian) Chen. I obtained a MS in Computer Science at National Taiwan University, advised by Prof. Hsin-Hsi Chen and a BS in Computer Science at National Yang Ming Chiao Tung University.

Nowadays, I am interested in (1) scaling test-time compute with generative verifiers, (2) building small yet performant LMs, and (3) improving reliability of LLM-based evaluation.

News

  • [06/2025] Started my Student Researcher internship at Google MTV! πŸš€
  • [06/2025] Check out our new preprint on negative RL for LLM reasoning! πŸ™Œ
  • [05/2025] Our AdaDecode is accepted at ICML 2025! πŸŽ‰
  • [04/2025] Check out our new preprint on analyzing LLM evaluators with verifiable benchmarks! πŸ™Œ
  • [01/2025] Our InstructRAG is accepted at ICLR 2025! πŸŽ‰
  • [09/2024] Our survey on persona is accepted at EMNLP 2024 findings! πŸŽ‰ Also checkout our accompanied repo!
  • [08/2024] Check out our new preprint on instructing LLMs for denoising RAG! πŸ™Œ
  • [07/2024] Our Traditional-Mandarian MMLU is accepted at the first COLM! πŸŽ‰
  • [04/2024] I will be starting my PhD journey at University of Virginia, advised by Prof. Yu Meng this fall! πŸš€
  • [10/2023] 3 papers (2 mian, 1 findings) are accepted at EMNLP 2023! πŸŽ‰
  • [06/2023] I graduated from National Taiwan University in June 2023. I am actively seeking full-time Ph.D. positions in 2024 Fall!

Selected Works

Preprint The Surprising Effectiveness of Negative Reinforcement in LLM Reasoning
[pdf] [x thread]
Xinyu Zhu, Mengzhou Xia, Zhepei Wei, Wei-Lin Chen, Danqi Chen, Yu Meng
Preprint Do LLM Evaluators Prefer Themselves for a Reason?
[pdf] [x thread]
Wei-Lin Chen, Zhepei Wei, Xinyu Zhu, Shi Feng, Yu Meng
ICLR 2025 InstructRAG: Instructing Retrieval-Augmented Generation via Self-Synthesized Rationales
[pdf] [repo]
Zhepei Wei, Wei-Lin Chen, Yu Meng
EMNLP 2024 Findings Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization
[pdf] [repo]
Yu-Min Tseng\(^*\), Yu-Chao Huang\(^*\), Teng-Yun Hsiao\(^*\),Wei-Lin Chen\(^*\), Chao-Wei Huang, Yu Meng, Yun-Nung Chen
EMNLP 2023 Self-ICL: Zero-Shot In-Context Learning with Self-Generated Demonstrations
[pdf] [poster]
Wei-Lin Chen\(^*\), Cheng-Kuang Wu\(^*\), Yun-Nung Chen, Hsin-Hsi Chen

Travels

  • [11/2024] EMNLP @ Florida πŸ‡ΊπŸ‡Έ
  • [10/2024] COLM @ Philadelphia πŸ‡ΊπŸ‡Έ
  • [12/2023] EMNLP @ Singapore πŸ‡ΈπŸ‡¬
  • [10/2023] CIKM @ Birmingham, UK πŸ‡¬πŸ‡§
  • [07/2023] SIGIR @ Taipei, Taiwan πŸ‡ΉπŸ‡Ό
  • [07/2023] ACL @ Toronto, Canada πŸ‡¨πŸ‡¦
  • [05/2023] ICLR @ Kigali, Rwanda πŸ‡·πŸ‡Ό
  • [10/2022] COLING @ Gyeongju, Korea πŸ‡°πŸ‡·