PeopleofSuperintelligence

Researcher

Bill Yuchen Lin

Also known as Yuchen Lin, Bill Lin, Bill Y. Lin

Bill Yuchen Lin is an AI researcher who works on language model alignment, reasoning and evaluation. He studied in Shanghai Jiao Tong University's IEEE Honor Class and completed his PhD at the University of Southern California in 2022. With collaborators, he developed CommonGen to test everyday commonsense reasoning and LLM-Blender to rank and combine answers from different language models. His 2024 work included RewardBench for evaluating reward models and WildBench for testing models on difficult queries from real users. Since November 2024, he is a member of technical staff at xAI and an affiliate assistant professor of electrical and computer engineering at the University of Washington.

Badges

Timeline

  1. 2014–2018

    Studies at Shanghai Jiao Tong

    He studies in the IEEE Honor Class at Shanghai Jiao Tong University from 2014 to 2018, completing his bachelor's degree in 2018.

  2. 2020

    CommonGen tests commonsense reasoning

    In 2020, he coauthors CommonGen, a task that asks models to turn a set of everyday concepts into a plausible sentence. Its dataset contains 77,000 commonsense descriptions covering 35,000 distinct sets of concepts.

  3. 2021

    RiddleSense tests creative language

    He coauthors RiddleSense in 2021, a multiple-choice task with about 5,700 examples of riddle-style commonsense questions. It tests reasoning about everyday knowledge and figurative language.

  4. 2022

    Doctorate and research internships

    He completes his computer science PhD at the University of Southern California in 2022, advised by Xiang Ren. Research internships take him to Microsoft Research Asia in 2017–2018, Google AI in 2019 with Sandeep Tata and in 2020 with William Cohen, and Meta's FAIR Lab in 2021 with Scott Yih.

  5. 2023

    Young investigator at Ai2

    In 2023, he is a postdoctoral young investigator at the Allen Institute for AI. He also serves on the advising committee of the TamingLLM workshop at SIGDIAL and INLG 2023.

  6. 2023

    Combining models with LLM-Blender

    With Dongfu Jiang and Xiang Ren, he coauthors LLM-Blender at ACL in 2023. The system compares pairs of answers from different language models, ranks the candidates, and combines the best-ranked answers into a new response.

  7. March 2024

    RewardBench evaluates reward models

    In March 2024, he coauthors RewardBench, a dataset and code base for evaluating reward models used to align language models with human preferences. It tests preferences between answers across chat, reasoning and safety tasks.

  8. June 2024

    WildBench uses real user queries

    He coauthors WildBench in June 2024, an evaluation framework built around difficult queries from real users. Its 1,024 tasks come from more than one million human–chatbot conversation logs, with task-specific checklists used to assess responses.

  9. Since November 2024

    Research at xAI and UW

    In November 2024, he joins xAI as a member of technical staff and becomes an affiliate assistant professor of electrical and computer engineering at the University of Washington. His Ai2 research scientist role spans 2023–2024, working with Yejin Choi. He continues at xAI and UW in October 2026, researching alignment, reasoning, coding agents, reward models and evaluation.

Career

  1. Since 2024xAIMember of Technical Staff
  2. Since 2024University of WashingtonAffiliate Assistant Professor of Electrical and Computer Engineering
  3. 2023–2024Allen Institute for AI (Ai2)Research Scientist
  4. 2021FAIR (Meta)Research Intern
  5. 2020GoogleResearch Intern
  6. 2019GoogleResearch Intern
  7. 2017–2018Microsoft Research AsiaResearch Intern
  8. Allen Institute for AI (Ai2)Postdoctoral Young Investigator

Education

  1. 2022University of Southern CaliforniaPhD in computer science
  2. 2018Shanghai Jiao Tong UniversityBachelor's degree

From the biography.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Bill Yuchen Lin yet. Send us a free-licensed photo.

  1. Bill Yuchen Lin | UW Department of Electrical & Computer Engineering ece.uw.edu. Checked 27 October 2026.
  2. CommonGen: A Constrained Text Generation Challenge for Generative Commonsense Reasoning - ACL Anthology aclanthology.org. Checked 27 October 2026.
  3. RiddleSense - RiddleSense | Project Site inklab.usc.edu. Checked 27 October 2026.
  4. bill@yuchenlin ~ $ yuchenlin.xyz. Checked 27 October 2026.
  5. TamingLLM Workshop @ SIGDIAL & INLG 2023 ctrlnlg.github.io. Checked 27 October 2026.
  6. LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusion - ACL Anthology aclanthology.org. Checked 27 October 2026.
  7. [2403.13787] RewardBench: Evaluating Reward Models for Language Modeling arxiv.org. Checked 27 October 2026.
  8. [2406.04770] WildBench: Benchmarking LLMs with Challenging Tasks from Real Users in the Wild arxiv.org. Checked 27 October 2026.