--- title: "Rowan Zellers — author profile and papers" description: "Rowan Zellers: 5 papers on Elamur Research (2019–2022), h-index 22 — affiliations, topics, co-authors and who cites them in our library." canonical: "https://research.elamur.ai/free/en/authors/rowan-zellers/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/rowan-zellers/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Rowan Zellers # Rowan Zellers Rowan Zellers works in natural language processing, with a focus on large neural language models: what they know about the world, how the text they produce can be evaluated, and how generation can be steered. Their papers here, from 2019 to 2022, include commonsense benchmarks, long-form text generators, measures that compare model output with human writing, decoding algorithms, and probing studies that link language representations to vision. Recurring themes are [text generation](https://md.research.elamur.ai/free/en/topics/text-generation/) and reasoning. The most cited of these papers is [PIQA](https://md.research.elamur.ai/free/en/papers/bisk-2019-piqa-reasoning-about-physical-commonsense/) (2019). It framed everyday physical knowledge as a question-answering task in which each goal comes with two candidate solutions. Crowdworkers took ideas from instructables.com guides and also wrote subtly flawed decoy answers. [Defending Against Neural Fake News](https://md.research.elamur.ai/free/en/papers/zellers-2019-defending-against-neural-fake-news/) (2019) borrowed threat modelling from computer security to study machine-written disinformation. It introduced Grover, a Transformer that writes whole news articles field by field, from domain and date through authors and headline to the body. [MAUVE](https://md.research.elamur.ai/free/en/papers/pillutla-2021-mauve-measuring-gap-between-neural/) (2021) proposed a way to measure how far a model's text distribution lies from human writing. It accounts both for output people would not produce and for plausible human text the model fails to cover. The newer work turns toward control and grounding. [NeuroLogic A*esque Decoding](https://md.research.elamur.ai/free/en/papers/lu-2022-neurologic-esque-decoding-constrained-text/) (2022) adapts ideas from A* search. It scores each candidate token with short lookahead continuations and an estimate of future cost, so that outputs can include or exclude chosen words. [Probing Contextual Language Models for Common Ground with Visual Representations](https://md.research.elamur.ai/free/en/papers/ilharco-2021-probing-contextual-language-models-common/) (2021) asked whether text-only embeddings of concrete nouns line up with object-detector features. The probe used for this was tested on categories it had not seen in training. Bylines on these papers list the [University of Washington](https://md.research.elamur.ai/free/en/universities/university-of-washington/) and the [Allen Institute for Artificial Intelligence](https://md.research.elamur.ai/free/en/labs/allen-institute-for-artificial-intelligence/) in Seattle, and later [OpenAI](https://md.research.elamur.ai/free/en/companies/openai/) in San Francisco, from 2021 to 2023. The most frequent co-author is [Yejin Choi](https://md.research.elamur.ai/free/en/authors/yejin-choi/). Others include [Yonatan Bisk](https://md.research.elamur.ai/free/en/authors/yonatan-bisk/), Sean Welleck, Ronan Le Bras and Ali Farhadi. Rowan Zellers has 5 papers in our library, in 2019–2022, most of them on [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/), [Text generation](https://md.research.elamur.ai/free/en/topics/text-generation/), [Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/); often with [Yejin Choi](https://md.research.elamur.ai/free/en/authors/yejin-choi/), [Ali Farhadi](https://md.research.elamur.ai/free/en/authors/ali-farhadi/), [Ronan Le Bras](https://md.research.elamur.ai/free/en/authors/ronan-le-bras/). 5 papers in our library · 2019–2022 [OpenAlex](https://openalex.org/A5030637764) · [ORCID](https://orcid.org/0000-0003-1361-9868) · [**h-index 22**](https://openalex.org/authors/A5030637764) · [38 works](https://openalex.org/works?filter=author.id:A5030637764) · [6 303 citations](https://openalex.org/works?filter=author.id:A5030637764&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [5 works in our library](https://md.research.elamur.ai/free/en/authors/rowan-zellers/#papers) · [4 052 citations in our library](https://md.research.elamur.ai/free/en/authors/rowan-zellers/cited-by/) · 2 more works in our library · [⭐ Paper Star rank #1149](https://md.research.elamur.ai/free/en/paper-stars/page/12/#star-1149) 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=Rowan+Zellers) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Rowan+Zellers) · [dblp](https://dblp.org/search?q=Rowan+Zellers) · [ACL Anthology](https://aclanthology.org/search/?q=Rowan+Zellers) ## Affiliations - [OpenAI (United States)](https://md.research.elamur.ai/free/en/companies/openai/) · San Francisco, United States —2021–2023 on 5 papers - [University of Washington](https://md.research.elamur.ai/free/en/universities/university-of-washington/) · Seattle, United States —2016–2022 on 25 papers - [Allen Institute for Artificial Intelligence](https://md.research.elamur.ai/free/en/labs/allen-institute-for-artificial-intelligence/) · Seattle, United States —2018–2022 on 10 papers - [University of Edinburgh](https://md.research.elamur.ai/free/en/universities/university-of-edinburgh/) · Edinburgh, United Kingdom —2021–2022 on 2 papers From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 2 more institutions on single bylines are not listed. ## Topics [Language models & pretraining 1](https://md.research.elamur.ai/free/en/topics/language-models/) · [Text generation 1](https://md.research.elamur.ai/free/en/topics/text-generation/) · [Reasoning 1](https://md.research.elamur.ai/free/en/topics/reasoning/) · [Interpretability 1](https://md.research.elamur.ai/free/en/topics/interpretability/) ## Works by year ### 2020s [**2022** 1 paper](https://md.research.elamur.ai/free/en/papers/lu-2022-neurologic-esque-decoding-constrained-text/) · [**2021** 2 papers](https://md.research.elamur.ai/free/en/authors/rowan-zellers/year/2021/) ### 2010s [**2019** 2 papers](https://md.research.elamur.ai/free/en/authors/rowan-zellers/year/2019/) ## Papers The most cited first (OpenAlex and our library). 1. **[PIQA: Reasoning about Physical Commonsense in Natural Language](https://md.research.elamur.ai/free/en/papers/bisk-2019-piqa-reasoning-about-physical-commonsense/)** *(2019 · 1 218 citations)* The paper asks whether language models trained only on text can handle everyday physical reasoning, such as choosing the right household object for a task. 2. **[Defending Against Neural Fake News](https://md.research.elamur.ai/free/en/papers/zellers-2019-defending-against-neural-fake-news/)** *(2019 · 461 citations)* Large text generators could let adversaries produce large amounts of fabricated news written to look like it comes from real outlets. 3. **[MAUVE: Measuring the Gap Between Neural Text and Human Text using Divergence Frontiers](https://md.research.elamur.ai/free/en/papers/pillutla-2021-mauve-measuring-gap-between-neural/)** *(2021 · 243 citations)* Judging how close machine-generated text is to human writing in open-ended generation is hard. The distributions involved are discrete and extremely high-dimensional, and they can only be reached through samples. The authors introduce MAUVE, a measure that compares a model's text distribution Q (language model plus… 4. **[NeuroLogic A*esque Decoding: Constrained Text Generation with Lookahead Heuristics](https://md.research.elamur.ai/free/en/papers/lu-2022-neurologic-esque-decoding-constrained-text/)** *(2022 · 136 citations)* Left-to-right decoders such as beam search and top-k sampling pick each token based only on the prefix generated so far. 5. **[Probing Contextual Language Models for Common Ground with Visual Representations](https://md.research.elamur.ai/free/en/papers/ilharco-2021-probing-contextual-language-models-common/)** *(2021 · 31 citations)* The paper asks how far text-only contextual language models encode information that lines up with visual representations of concrete nouns, without any visual supervision during pre-training. ## Also in our library 2 more works of theirs are in our library without a page here yet; they will appear as we publish them. ## Co-authors here 12 of 21 co-authors with a page, the most papers together first. - [Yejin Choi](https://md.research.elamur.ai/free/en/authors/yejin-choi/) ×4 - [Ali Farhadi](https://md.research.elamur.ai/free/en/authors/ali-farhadi/) ×2 - [Ronan Le Bras](https://md.research.elamur.ai/free/en/authors/ronan-le-bras/) ×2 - [Sean Welleck](https://md.research.elamur.ai/free/en/authors/sean-welleck/) ×2 - [Yonatan Bisk](https://md.research.elamur.ai/free/en/authors/yonatan-bisk/) ×2 - [Ari Holtzman](https://md.research.elamur.ai/free/en/authors/ari-holtzman/) - [Daniel Khashabi](https://md.research.elamur.ai/free/en/authors/daniel-khashabi/) - [Gabriel Ilharco](https://md.research.elamur.ai/free/en/authors/gabriel-ilharco/) - [Hannah Rashkin](https://md.research.elamur.ai/free/en/authors/hannah-rashkin/) - [Hannaneh Hajishirzi](https://md.research.elamur.ai/free/en/authors/hannaneh-hajishirzi/) - [Jianfeng Gao](https://md.research.elamur.ai/free/en/authors/jianfeng-gao/) - [John Thickstun](https://md.research.elamur.ai/free/en/authors/john-thickstun/) ## Cited by Papers of our library cite their works 4 052 times (of the 140 884 papers whose bibliographies we read); [226 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/rowan-zellers/cited-by/).