PeopleofSuperintelligence

Researcher

Yanda Chen

Yanda Chen is a language-model researcher whose work spans in-context learning, model explanations, and AI safety. Early studies examined language bias and cross-language retrieval, while later papers proposed methods for few-shot adaptation and tested whether explanations predict model behavior. They completed a computer-science PhD at Columbia in 2024 and coauthored Anthropic safety research on reasoning-trace faithfulness and pretraining-data filtering. As of October 2026, they are a Research Scientist on Anthropic's Code RL team, where they train Claude models toward stronger coding and agentic capabilities.

Now at
Anthropic
Badges

Timeline

  1. 2019–2021

    Bias research and cross-language NLP

    Chen coauthors a 2019 study that tests bias in models used to flag aggression in gang-related posts. In summer 2020 and summer 2021, they intern at IBM Research and AWS AI; in April 2021, they earn a computer-science bachelor's degree at Columbia University. Their 2021 ACL paper develops cross-language sentence selection for low-resource languages with data augmentation and rationale training, and the Computing Research Association gives the work an honorable mention.

  2. 2022

    In-context tuning at ACL

    Chen and four coauthors introduce in-context tuning in an ACL 2022 paper. The method recasts task adaptation and prediction as sequence prediction from instructions, labeled examples, and target inputs.

  3. 2023–2024

    A Microsoft internship and ICML work

    In summer 2023, Chen interns at Microsoft Research under Chandan Singh and Xiaodong Liu. With seven coauthors, they post Do Models Explain Themselves? in 2023, proposing counterfactual simulatability to test whether explanations help people infer model outputs on changed inputs. The ICML 2024 paper finds low precision in LLM explanations, with precision uncorrelated with plausibility.

  4. 2024

    A Columbia PhD and ICL study

    Chen completes a Columbia computer-science PhD in summer 2024, co-advised by Kathy McKeown, He He, and Zhou Yu. An ACL 2024 paper links language-model in-context learning to parallel phrase structures in pretraining data and finds that removing them cuts ICL accuracy by 51%.

  5. 2025

    Faithfulness and data filtering

    Chen coauthors an Anthropic study in 2025 that finds chain-of-thought traces reveal hint use in fewer than 20% of examples in most tested settings. Another 2025 paper filters harmful material from pretraining data and cuts model performance on a harmful-capabilities evaluation by 33% relative to a random baseline.

  6. As of October 2026

    Code RL at Anthropic

    As of October 2026, Chen is a Member of Technical Staff (Research Scientist) on Anthropic's Code RL team, where they train Claude models toward stronger coding and agentic capabilities.

Career

  1. NowAnthropicResearch Scientist
  2. 2023Microsoft ResearchIntern
  3. 2021AmazonIntern
  4. 2020IBM ResearchIntern

Education

  1. 2021Columbia UniversityBachelor's degree in computer science

From the biography.

  1. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Yanda Chen. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  2. 22 of 133 people shaping AI with a former OpenAI role now work at AnthropicIncludes Anthropic. Our estimates of where OpenAI recruits came from and where 133 people with former OpenAI roles work now, including the companies and labs they founded.
  3. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Anthropic. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  4. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Anthropic. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Yanda Chen yet. Send us a free-licensed photo.

  1. Detecting and Reducing Bias in a High Stakes Domain aclanthology.org. Checked 20 October 2026.
  2. Yanda Chen yandachen.github.io. Checked 20 October 2026.
  3. Cross-language Sentence Selection via Data Augmentation and Rationale Training aclanthology.org. Checked 20 October 2026.
  4. Three CS Students Recognized By The Computing Research Association cs.columbia.edu. Checked 20 October 2026.
  5. Meta-learning via Language Model In-context Tuning aclanthology.org. Checked 20 October 2026.
  6. Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations arxiv.org. Checked 20 October 2026.
  7. Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations proceedings.mlr.press. Checked 20 October 2026.
  8. Parallel Structures in Pre-training Data Yield In-Context Learning aclanthology.org. Checked 20 October 2026.
  9. Reasoning Models Don't Always Say What They Think arxiv.org. Checked 20 October 2026.
  10. Enhancing Model Safety through Pretraining Data Filtering alignment.anthropic.com. Checked 20 October 2026.