--- title: "Rishi Bommasani — author profile and papers" description: "Rishi Bommasani, Demos: 3 papers on Elamur Research (2020–2022), h-index 16 — affiliations, topics, co-authors and who cites them in our library." canonical: "https://research.elamur.ai/free/en/authors/rishi-bommasani/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/rishi-bommasani/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Rishi Bommasani # Rishi Bommasani Rishi Bommasani works on large pretrained language models. The work asks two kinds of question: what the internal representations of these models encode, and what such models can and cannot do as they grow larger and are adopted more widely. Their papers here date from 2020 to 2022 and favour analysis over building new systems. They adapt existing interpretation methods, re-examine published results across many models, and take stock of a fast-moving area, all within [language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/). The most cited of these papers is [On the Opportunities and Risks of Foundation Models](https://md.research.elamur.ai/free/en/papers/bommasani-2021-opportunities-risks-foundation-models/) (2021). This lengthy survey came out of CRFM, a Stanford centre devoted to the subject, and was written by more than a hundred contributors from many disciplines. The report introduced the term "foundation models" for large self-supervised models that are later adapted to many downstream uses, such as BERT, GPT-3, CLIP and DALL-E. It surveys their capabilities, their applications in areas like healthcare and law, and the concerns they raise. An earlier paper, [Interpreting Pretrained Contextualized Representations via Reductions to Static Embeddings](https://md.research.elamur.ai/free/en/papers/bommasani-2020-interpreting-pretrained-contextualized-representations-reductions/) (2020), proposed a parameter-free way to collapse a contextual encoder like BERT into a plain word lookup table. It does this by pooling subword vectors and then pooling across many sampled sentences, so that tools designed for static word vectors can be applied to contextual models. The newest work here, [Emergent Abilities of Large Language Models](https://md.research.elamur.ai/free/en/papers/wei-2022-emergent-abilities-large-language-models/) (2022), turns to the question of scale. Drawing on reported results for GPT-3, LaMDA, Gopher, Chinchilla and PaLM rather than training new models, it points to tasks where smaller models perform no better than chance and larger ones improve abruptly past some size. Such abilities cannot be predicted simply by extrapolating how smaller models behave, and most of the examples involve few-shot prompting. Bommasani is based at [Stanford University](https://md.research.elamur.ai/free/en/universities/stanford-university/), with bylines there from 2020 onward, following earlier papers at [Cornell University](https://md.research.elamur.ai/free/en/universities/cornell-university/) in 2019 and 2020. Co-authors include [Claire Cardie](https://md.research.elamur.ai/free/en/authors/claire-cardie/), [Michael S. Bernstein](https://md.research.elamur.ai/free/en/authors/michael-s-bernstein/) and [Jason Wei](https://md.research.elamur.ai/free/en/authors/jason-wei/), along with others on the emergent-abilities and foundation-model papers. Rishi Bommasani has 3 papers in our library, in 2020–2022, most of them on [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/); often with [Antoine Bosselut](https://md.research.elamur.ai/free/en/authors/antoine-bosselut/), [Barret Zoph](https://md.research.elamur.ai/free/en/authors/barret-zoph/), [Claire Cardie](https://md.research.elamur.ai/free/en/authors/claire-cardie/). 3 papers in our library · 2020–2022 current affiliation [Demos](https://ror.org/032bmj362) · New York United States; [Demos](https://ror.org/00048tr39) · London United Kingdom; [Massachusetts Institute of Technology](https://md.research.elamur.ai/free/en/universities/massachusetts-institute-of-technology/) · Cambridge United States; [Stanford University](https://md.research.elamur.ai/free/en/universities/stanford-university/) · Stanford United States [OpenAlex](https://openalex.org/A5069576651) · [ORCID](https://orcid.org/0000-0002-9616-5138) · [Wikidata](https://www.wikidata.org/wiki/Q133266101) · [**h-index 16**](https://openalex.org/authors/A5069576651) · [80 works](https://openalex.org/works?filter=author.id:A5069576651) · [5,356 citations](https://openalex.org/works?filter=author.id:A5069576651&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [3 works in our library](https://md.research.elamur.ai/free/en/authors/rishi-bommasani/#papers) · [2,198 citations in our library](https://md.research.elamur.ai/free/en/authors/rishi-bommasani/cited-by/) · 1 more work in our library · [⭐ Paper Star rank #469](https://md.research.elamur.ai/free/en/paper-stars/page/5/#star-469) The profile is the library's match of the name; two people with one name are not always told apart. Elsewhere: [Google Scholar](https://scholar.google.com/citations?view_op=search_authors&mauthors=Rishi+Bommasani) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Rishi+Bommasani) · [dblp](https://dblp.org/search?q=Rishi+Bommasani) · [ACL Anthology](https://aclanthology.org/search/?q=Rishi+Bommasani) ## Affiliations [🎓**Stanford University** Stanford · United States 2020–2026 · 42 papers*current*](https://md.research.elamur.ai/free/en/universities/stanford-university/) · [🤝**Demos** New York · United States 2026 · 1 paper*current*](https://ror.org/032bmj362) · [🤝**Demos** London · United Kingdom 2026 · 1 paper*current*](https://ror.org/00048tr39) · [🎓**Massachusetts Institute of Technology** Cambridge · United States 2026 · 1 paper*current*](https://md.research.elamur.ai/free/en/universities/massachusetts-institute-of-technology/) · [🎓**Cornell University** Ithaca · United States 2019–2020 · 2 papers](https://md.research.elamur.ai/free/en/universities/cornell-university/) From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 3 more institutions on single bylines are not listed. ## Topics [Language models & pretraining 2](https://md.research.elamur.ai/free/en/topics/language-models/) ## Works by year ### 2020s [**2022** 1 paper](https://md.research.elamur.ai/free/en/papers/wei-2022-emergent-abilities-large-language-models/) · [**2021** 1 paper](https://md.research.elamur.ai/free/en/papers/bommasani-2021-opportunities-risks-foundation-models/) · [**2020** 1 paper](https://md.research.elamur.ai/free/en/papers/bommasani-2020-interpreting-pretrained-contextualized-representations-reductions/) ## Papers The most cited first (OpenAlex and our library). 1. **[On the Opportunities and Risks of Foundation Models](https://md.research.elamur.ai/free/en/papers/bommasani-2021-opportunities-risks-foundation-models/)** *(2021 · 3,044 citations)* This report from Stanford's Center for Research on Foundation Models (CRFM) names and examines a new way of building AI. 2. **[Emergent Abilities of Large Language Models](https://md.research.elamur.ai/free/en/papers/wei-2022-emergent-abilities-large-language-models/)** *(2022 · 2,215 citations)* Scaling laws usually describe language model performance improving smoothly as models grow. This paper looks at the cases where that doesn't happen. The authors call an ability "emergent" if small models lack it and larger ones have it, which means you can't forecast it by extending the trend from smaller models.… 3. **[Interpreting Pretrained Contextualized Representations via Reductions to Static Embeddings](https://md.research.elamur.ai/free/en/papers/bommasani-2020-interpreting-pretrained-contextualized-representations-reductions/)** *(2020 · 248 citations)* Tools for interpreting static word vectors are mature, yet they are rarely used to study contextual encoders such as BERT. ## Also in our library 1 more work of theirs is in our library without a page here yet; it will appear as we publish them. ## Co-authors here 12 of 16 co-authors with a page, the most papers together first. - [Antoine Bosselut](https://md.research.elamur.ai/free/en/authors/antoine-bosselut/) - [Barret Zoph](https://md.research.elamur.ai/free/en/authors/barret-zoph/) - [Claire Cardie](https://md.research.elamur.ai/free/en/authors/claire-cardie/) - [Colin Raffel](https://md.research.elamur.ai/free/en/authors/colin-raffel/) - [Dani Yogatama](https://md.research.elamur.ai/free/en/authors/dani-yogatama/) - [Denny Zhou](https://md.research.elamur.ai/free/en/authors/denny-zhou/) - [Donald Metzler](https://md.research.elamur.ai/free/en/authors/donald-metzler/) - [Drew A. Hudson](https://md.research.elamur.ai/free/en/authors/drew-a-hudson/) - [Jason Wei](https://md.research.elamur.ai/free/en/authors/jason-wei/) - [Kelly Davis](https://md.research.elamur.ai/free/en/authors/kelly-davis/) - [Maarten Bosma](https://md.research.elamur.ai/free/en/authors/maarten-bosma/) - [Michael S. Bernstein](https://md.research.elamur.ai/free/en/authors/michael-s-bernstein/) ## Cited by Papers of our library cite their works 2,198 times (of the 140,884 papers whose bibliographies we read); [196 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/rishi-bommasani/cited-by/).