Builders & open source
Yuanzhi Li
First author of the phi-1.5 report and co-author of TinyStories and Microsoft's other Phi model reports
Yuanzhi Li is first author of the 2023 phi-1.5 technical report, which describes a 1.3 billion parameter model trained largely on synthetic "textbook quality" data and states that the model was open-sourced; its Hugging Face model card lists the MIT license. Li co-authored the TinyStories paper with Ronen Eldan, with both listed at Microsoft Research, and is a co-author of the phi-1, Phi-3, and Phi-4 technical reports and a listed contributor to Phi-2.1234567
At a glance
- Current role
- None confirmed in the sources
- Fields
- Language models, Datasets and data
- Work listed
- 2
- Sources cited
- 8
- Profile reviewed
Research
Li's recent work centers on small language models trained on carefully generated data. TinyStories (2023), written with Ronen Eldan, built a synthetic dataset of short stories using only words a young child would understand and showed that models with fewer than 10 million parameters trained on it could produce fluent, consistent English.3
Li is also a co-author of "Sparks of Artificial General Intelligence: Early experiments with GPT-4" (2023), which examined an early version of GPT-4, while it was still in development, across tasks in mathematics, coding, vision, medicine, law, and psychology.8
Phi models
Li is a co-author of the phi-1 report "Textbooks Are All You Need" (2023), on a code model trained on textbook-quality data, and first author of the phi-1.5 report, which extended the approach to common sense reasoning in natural language with a 1.3-billion-parameter model. Li also co-authored the Phi-3 and Phi-4 technical reports and is listed among the contributors to Phi-2.412567
Work and contributions
Phi small language models Open weights
First author of the phi-1.5 technical report (2023), co-author of the phi-1 (2023), Phi-3 (2024), and Phi-4 (2024) reports, and a listed contributor on the Microsoft Research blog post announcing Phi-2. The phi-1.5 model card on Hugging Face lists the MIT license.145672
TinyStories Published research
Co-author with Ronen Eldan of the 2023 paper introducing TinyStories, a synthetic dataset of short stories generated by GPT-3.5 and GPT-4, used to train and evaluate language models with fewer than 10 million parameters.3
Timeline
- Co-authored "Sparks of Artificial General Intelligence", a study of an early version of GPT-4.8
- Released the TinyStories paper with Ronen Eldan.3
- Co-authored "Textbooks Are All You Need", the phi-1 technical report.4
- Released the phi-1.5 technical report as first author.1
- Listed among the contributors in Microsoft Research's announcement of Phi-2.7
- Co-authored the Phi-3 technical report.5
- Co-authored the Phi-4 technical report.6
Documented roles
The sources do not confirm a current role; these are the roles they document.
- Listed with a Microsoft Research affiliation on the 2023 TinyStories paper, Microsoft Research3
Companies in their history
- Microsoft 75 people in the wiki
Related people
- Luca SoldainiConnected to Microsoft
- Ronen EldanConnected to Microsoft
- Bill DolanConnected to Microsoft
- Rangan MajumderConnected to Microsoft
- David TalbyConnected to Microsoft
- Oren EtzioniDatasets and data · Language models
- Connor LeahyLanguage models · Datasets and data
- Marina DanilevskyLanguage models · Datasets and data
- George HalalLanguage models · Datasets and data
- Albert JiangLanguage models · Datasets and data
Suggest a correction or removal
Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.
References
8 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026
- Textbooks Are All You Need II: phi-1.5 technical report (arXiv 2309.05463)
- microsoft/phi-1_5 model card (license MIT)
- TinyStories: How Small Can Language Models Be and Still Speak Coherent English? (arXiv 2305.07759)
- Textbooks Are All You Need (arXiv 2306.11644)
- Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone (arXiv 2404.14219)
- Phi-4 Technical Report (arXiv 2412.08905)
- Phi-2: The surprising power of small language models
- Sparks of Artificial General Intelligence: Early experiments with GPT-4 (arXiv 2303.12712)