
Academic lab · Princeton, United States
Princeton NLP Group
The Princeton NLP group, source of SWE-bench, SimPO and Tree of Thoughts, with many open benchmarks, methods and models.
Princeton, United States
- Type
- Academic lab
- Based in
- Princeton, United States
- Key papers
- 4
- Featured projects
- 7
About
Princeton NLP is the natural language processing research group at Princeton University. Its GitHub organization hosts widely used research code and its Hugging Face organization hosts models and datasets.
The group's SWE-bench benchmark, published at ICLR 2024, tests whether language models can resolve real GitHub issues and became a standard evaluation for coding agents. SimPO, presented at NeurIPS 2024, is a reference-free preference optimization method, and Tree of Thoughts (NeurIPS 2023) is a deliberate problem-solving framework for language models.
Karthik Narasimhan is among the authors of SWE-bench; Danqi Chen also leads work in the group.
Mission and ambitionsThe group studies how language models learn, reason and act, and releases the resulting benchmarks, methods and code so that other researchers can reproduce and extend them.
Known for
- 01
SWE-bench
- 02
SimPO preference optimization
- 03
Tree of Thoughts
- 04
Open academic NLP research
Milestones
- 2023
Tree of Thoughts
Introduced deliberate problem solving with language models.
- 2023
SWE-bench
Released the benchmark of 2,294 real GitHub issue tasks.
- 2024
SimPO
Released a simple reference-free preference optimization method with models.
Open source
Tree of Thoughts
Contrastive sentence embeddings
Memory-efficient fine-tuning
Simple preference optimisation
- SWE-bench_Verifieddataset
Verified SWE-bench subset
- SWE-benchdataset
SWE-bench
Long-context model
Research papers
Notable releases
Openness
MIT-licensed code; benchmarks and models on Hugging Face.
People
- Karthik NarasimhanAssociate professor, co-author of SWE-bench and Tree of Thoughts
Leads work on language agents.
- Danqi ChenAssociate professor, co-leader of the Princeton NLP Group
Also associate director of Princeton Language and Intelligence.
- Shunyu YaoFirst author of Tree of Thoughts, formerly a Princeton PhD student
Now a researcher at OpenAI.
- John YangCreator of SWE-bench and SWE-agent
Earlier work done at Princeton; now a PhD student at Stanford.
Listed from the lab's own pages and public profiles.
Contact and community
Via website.
Sources
Compiled from the lab's public pages, Hugging Face and GitHub. Last checked 2026-11-17. Something wrong or out of date? Tell us on Discord.



