--- title: "Susan Zhang — author profile and papers" description: "Susan Zhang, La Trobe University: 2 papers on Elamur Research (2022–2023), h-index 5 — affiliations, topics, co-authors and who cites them in our library." canonical: "https://research.elamur.ai/free/en/authors/susan-zhang/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/susan-zhang/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Susan Zhang # Susan Zhang Susan Zhang works on large language models and how they are trained. The papers here cover [language models and pretraining](https://md.research.elamur.ai/free/en/topics/language-models/), especially building very large autoregressive transformers and measuring how their behaviour changes with scale. The methods are empirical: models are trained across a wide range of sizes, and the outcomes are compared to find out what changes as models grow. The most cited of these papers is [OPT: Open Pre-trained Transformer Language Models](https://md.research.elamur.ai/free/en/papers/zhang-2022-opt-open-pre-trained-transformer/) (2022). It released a family of decoder-only transformers, from small to very large, whose design closely follows GPT-3. The aim was to let researchers outside a few well-funded labs study models of this kind directly, with access to the weights. The paper describes how the pretraining corpus was put together and deduplicated. It also gives a frank account of the hardware failures and other difficulties of training the largest model. The newer work moves from text-only models to models that handle several kinds of input. [Scaling Laws for Generative Mixed-Modal Language Models](https://md.research.elamur.ai/free/en/papers/aghajanyan-2023-scaling-laws-generative-mixed-modal/) (2023) turns images, speech, code, molecules and text into discrete tokens for a single generative model. It then asks whether greater scale can offset the interference between modalities that share one network. Zhang is currently at La Trobe University and earlier appears with OpenAI (United States), linked as [OpenAI](https://md.research.elamur.ai/free/en/companies/openai/). Frequent co-authors on these papers include [Naman Goyal](https://md.research.elamur.ai/free/en/authors/naman-goyal/) and [Stephen Roller](https://md.research.elamur.ai/free/en/authors/stephen-roller/). Other collaborators include [Armen Aghajanyan](https://md.research.elamur.ai/free/en/authors/armen-aghajanyan/) and [Luke Zettlemoyer](https://md.research.elamur.ai/free/en/authors/luke-zettlemoyer/). Susan Zhang has 2 papers in our library, in 2022–2023, most of them on [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/); often with [Naman Goyal](https://md.research.elamur.ai/free/en/authors/naman-goyal/), [Stephen Roller](https://md.research.elamur.ai/free/en/authors/stephen-roller/), [Alexis Conneau](https://md.research.elamur.ai/free/en/authors/alexis-conneau/). 2 papers in our library · 2022–2023 [OpenAlex](https://openalex.org/A5051860831) · [ORCID](https://orcid.org/0000-0003-4939-3944) · [**h-index 5**](https://openalex.org/authors/A5051860831) · [231 works](https://openalex.org/works?filter=author.id:A5051860831) · [1,005 citations](https://openalex.org/works?filter=author.id:A5051860831&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [📄 2 works in our library](https://md.research.elamur.ai/free/en/authors/susan-zhang/#papers) · [🔗 1,741 citations in our library](https://md.research.elamur.ai/free/en/authors/susan-zhang/cited-by/) · [⭐ Paper Star rank #1328](https://md.research.elamur.ai/free/en/paper-stars/page/14/#star-1328) 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=Susan+Zhang) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Susan+Zhang) · [dblp](https://dblp.org/search?q=Susan+Zhang) · [ACL Anthology](https://aclanthology.org/search/?q=Susan+Zhang) ## Affiliations [🎓**La Trobe University** Melbourne · Australia 2025 · 1 paper*current*](https://ror.org/01rxfrp27) · [🎓**University of Wollongong** Wollongong · Australia 2021–2025 · 8 papers](https://ror.org/00jtmb277) · [🏢**OpenAI** San Francisco · United States 2019 · 2 papers](https://md.research.elamur.ai/free/en/companies/openai/) From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 4 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 [**2023** 1 paper](https://md.research.elamur.ai/free/en/papers/aghajanyan-2023-scaling-laws-generative-mixed-modal/) · [**2022** 1 paper](https://md.research.elamur.ai/free/en/papers/zhang-2022-opt-open-pre-trained-transformer/) ## Papers The most cited first (OpenAlex and our library). 1. **[OPT: Open Pre-trained Transformer Language Models](https://md.research.elamur.ai/free/en/papers/zhang-2022-opt-open-pre-trained-transformer/)** *(2022 · 1,709 citations)* The largest language models are very expensive to train, so only a few labs can reproduce them. Models available through paid APIs do not provide access to their weights, which makes them hard to study. The authors from Meta AI created OPT, a suite of autoregressive decoder-only transformers ranging from 125M to 175B… 2. **[Scaling Laws for Generative Mixed-Modal Language Models](https://md.research.elamur.ai/free/en/papers/aghajanyan-2023-scaling-laws-generative-mixed-modal/)** *(2023 · 39 citations)* Single generative models can treat images, speech, code and text as token streams. It is unclear, however, whether added scale overcomes the interference that arises when several modalities share one network. The authors ran more than 250 training experiments covering seven modalities: Text, Image, Image-Text… ## Co-authors here 12 of 16 co-authors with a page, the most papers together first. - [Naman Goyal](https://md.research.elamur.ai/free/en/authors/naman-goyal/) ×2 - [Stephen Roller](https://md.research.elamur.ai/free/en/authors/stephen-roller/) ×2 - [Alexis Conneau](https://md.research.elamur.ai/free/en/authors/alexis-conneau/) - [Armen Aghajanyan](https://md.research.elamur.ai/free/en/authors/armen-aghajanyan/) - [Karen Hambardzumyan](https://md.research.elamur.ai/free/en/authors/karen-hambardzumyan/) - [Luke Zettlemoyer](https://md.research.elamur.ai/free/en/authors/luke-zettlemoyer/) - [Mikel Artetxe](https://md.research.elamur.ai/free/en/authors/mikel-artetxe/) - [Mona Diab](https://md.research.elamur.ai/free/en/authors/mona-diab/) - [Moya Chen](https://md.research.elamur.ai/free/en/authors/moya-chen/) - [Myle Ott](https://md.research.elamur.ai/free/en/authors/myle-ott/) - [Omer Levy](https://md.research.elamur.ai/free/en/authors/omer-levy/) - [Shuohui Chen](https://md.research.elamur.ai/free/en/authors/shuohui-chen/) ## Cited by Papers of our library cite their works 1,741 times (of the 140,884 papers whose bibliographies we read); [166 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/susan-zhang/cited-by/).