--- title: "Michael Pieler — author profile and papers" description: "Michael Pieler, Edison Scientific Inc.: 3 papers on Elamur Research (2022–2024), h-index 10 — affiliations, topics, co-authors and who cites them in our…" canonical: "https://research.elamur.ai/free/en/authors/michael-pieler/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/michael-pieler/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Michael Pieler # Michael Pieler Michael Pieler works on large language models, with papers here that span 2022 to 2024. Two questions run through them: how to build and openly release autoregressive Transformer models, and how to adapt an existing model to a language it handles poorly. The methods are mostly about pretraining: choosing and cleaning corpora, balancing data sources and evaluating the resulting models. This places the work under [language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) and, in its more recent part, [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/). The most cited paper is [GPT-NeoX-20B](https://md.research.elamur.ai/free/en/papers/black-2022-gpt-neox-20b-open-source/) (2022). It presents a dense Transformer of the decoder-only kind with 20 billion parameters, which EleutherAI trained on the Pile corpus. The weights were made freely downloadable so that people without access to commercial systems could examine how large models behave, for instance in work on safety, interpretability or scaling. When the paper was submitted, no larger dense autoregressive model was openly available in this way. The newer work turns from building a model from scratch to adapting one to a new language. [Arabic Stable LM](https://md.research.elamur.ai/free/en/papers/alyafeai-2024-arabic-stable-lm-adapting-stable/) (2024) continues pretraining the 1.6B-parameter Stable LM 2 on English mixed with cleaned Arabic web, news and e-book text. A chat version adds synthetic Arabic dialogues that an LLM produced by rephrasing existing material. The authors measure how heavily the tokenizer splits Arabic words and correct the sampling mix for the token counts this inflates. They also re-derive quality-filter thresholds from statistics of Arabic itself. The paper reports that these compact models match or outperform much larger Arabic-focused ones. Pieler is currently at Edison Scientific Inc. in San Francisco, after an earlier affiliation with [New York University](https://md.research.elamur.ai/free/en/universities/new-york-university/) in 2023. Co-authors on the papers here include [Zaid Alyafeai](https://md.research.elamur.ai/free/en/authors/zaid-alyafeai/), first author of the Arabic work, as well as [Stella Biderman](https://md.research.elamur.ai/free/en/authors/stella-biderman/), [Leo Gao](https://md.research.elamur.ai/free/en/authors/leo-gao/) and [Jason Phang](https://md.research.elamur.ai/free/en/authors/jason-phang/). Michael Pieler has 3 papers in our library, in 2022–2024, most of them on [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/), [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/); often with [Aaron Mueller](https://md.research.elamur.ai/free/en/authors/aaron-mueller/), [Alexis Ross](https://md.research.elamur.ai/free/en/authors/alexis-ross/), [Alicia Parrish](https://md.research.elamur.ai/free/en/authors/alicia-parrish/). 3 papers in our library · 2022–2024 [OpenAlex](https://openalex.org/A5029968236) · [ORCID](https://orcid.org/0000-0001-9186-7045) · [**h-index 10**](https://openalex.org/authors/A5029968236) · [36 works](https://openalex.org/works?filter=author.id:A5029968236) · [987 citations](https://openalex.org/works?filter=author.id:A5029968236&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [📄 3 works in our library](https://md.research.elamur.ai/free/en/authors/michael-pieler/#papers) · [🔗 532 citations in our library](https://md.research.elamur.ai/free/en/authors/michael-pieler/cited-by/) · [1 more in our library without a page yet](#also) · [⭐ Paper Star rank #2104](https://md.research.elamur.ai/free/en/paper-stars/page/22/#star-2104) 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=Michael+Pieler) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Michael+Pieler) · [dblp](https://dblp.org/search?q=Michael+Pieler) · [ACL Anthology](https://aclanthology.org/search/?q=Michael+Pieler) ## Affiliations [🏢**Edison Scientific Inc.** San Francisco · United States 2025 · 1 paper*current*](https://ror.org/01p9amb16) · [🔬**Max Planck Institute for Dynamics of Complex Technical Systems** Magdeburg · Germany 2014–2021 · 20 papers](https://ror.org/030h7k016) · [🔬**Max Planck Society** Munich · Germany 2014–2016 · 13 papers](https://md.research.elamur.ai/free/en/labs/max-planck-society/) · [🎓**New York University** New York · United States 2023 · 2 papers](https://md.research.elamur.ai/free/en/universities/new-york-university/) From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. ## Topics [Language models & pretraining (1)](https://md.research.elamur.ai/free/en/topics/language-models/) · [Semitic languages (1)](https://md.research.elamur.ai/free/en/topics/semitic-languages/) ## Works by year ### 2020s [**2024** 1 paper](https://md.research.elamur.ai/free/en/papers/alyafeai-2024-arabic-stable-lm-adapting-stable/) · [**2023** 1 paper](https://md.research.elamur.ai/free/en/papers/mckenzie-2023-inverse-scaling-bigger-isn-t/) · [**2022** 1 paper](https://md.research.elamur.ai/free/en/papers/black-2022-gpt-neox-20b-open-source/) ## Papers The most cited first (OpenAlex and our library). 1. **[GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://md.research.elamur.ai/free/en/papers/black-2022-gpt-neox-20b-open-source/)** *(2022 · 818 citations)* Most large language models are proprietary: they sit behind commercial APIs or are not shared at all. This makes it hard for outside researchers to study behaviour that only appears at scale, such as in safety, interpretability and scaling research. EleutherAI addresses this gap with GPT-NeoX-20B, a… 2. **[Inverse Scaling: When Bigger Isn't Better](https://md.research.elamur.ai/free/en/papers/mckenzie-2023-inverse-scaling-bigger-isn-t/)** *(2023 · 99 citations)* Scaling laws predict that larger language models steadily lower training loss and usually improve on downstream tasks. 3. **[Arabic Stable LM: Adapting Stable LM 2 1.6B to Arabic](https://md.research.elamur.ai/free/en/papers/alyafeai-2024-arabic-stable-lm-adapting-stable/)** *(2024 · 6 citations)* Most Arabic-focused LLMs have at least 7B parameters, which makes them costly to run and slow at inference. ## 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 co-authors with a page, the most papers together first. - [Aaron Mueller](https://md.research.elamur.ai/free/en/authors/aaron-mueller/) - [Alexis Ross](https://md.research.elamur.ai/free/en/authors/alexis-ross/) - [Alicia Parrish](https://md.research.elamur.ai/free/en/authors/alicia-parrish/) - [Alisa Liu](https://md.research.elamur.ai/free/en/authors/alisa-liu/) - [Duy Phung](https://md.research.elamur.ai/free/en/authors/duy-phung/) - [Horace He](https://md.research.elamur.ai/free/en/authors/horace-he/) - [Jason Phang](https://md.research.elamur.ai/free/en/authors/jason-phang/) - [Kyle McDonell](https://md.research.elamur.ai/free/en/authors/kyle-mcdonell/) - [Leo Gao](https://md.research.elamur.ai/free/en/authors/leo-gao/) - [Quentin Anthony](https://md.research.elamur.ai/free/en/authors/quentin-anthony/) - [Stella Biderman](https://md.research.elamur.ai/free/en/authors/stella-biderman/) - [Zaid Alyafeai](https://md.research.elamur.ai/free/en/authors/zaid-alyafeai/) ## Cited by Papers of our library cite their works 532 times (of the 140,884 papers whose bibliographies we read); [61 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/michael-pieler/cited-by/).