--- title: "Karina Nguyen — author profile and papers" description: "Karina Nguyen, Johnson Center for Child Health and Development: 3 papers on Elamur Research (2022–2023), h-index 8 — affiliations, topics, co-authors and…" canonical: "https://research.elamur.ai/free/en/authors/karina-nguyen/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/karina-nguyen/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Karina Nguyen # Karina Nguyen Karina Nguyen works on evaluating large language models. The papers here ask what such models do, which views they express, and whether the reasoning they write out reflects how they arrive at an answer. They fall under [language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/), [evaluation & benchmarks](https://md.research.elamur.ai/free/en/topics/evaluation-benchmarks/), and reasoning. The usual method is to build purpose-made test collections or controlled interventions and apply them to assistant models tuned with human feedback. The most cited of the papers here is [Discovering Language Model Behaviors with Model-Written Evaluations](https://md.research.elamur.ai/free/en/papers/perez-2022-discovering-language-model-behaviors-model/) (2022). In that work one model generates test items for a target behaviour and a second model screens them for consistency with their labels. The pipelines use different amounts of human effort and produced 154 evaluation sets. [Towards Measuring the Representation of Subjective Global Opinions in Language Models](https://md.research.elamur.ai/free/en/papers/durmus-2023-towards-measuring-representation-subjective-global/) (2023) turns questions from the Pew Global Attitudes survey and the World Values Survey into multiple-choice items. It then compares a dialogue model's probabilities over the answers with how respondents in various countries replied. The 2023 work also asks whether step-by-step explanations can be trusted. [Measuring Faithfulness in Chain-of-Thought Reasoning](https://md.research.elamur.ai/free/en/papers/lanham-2023-measuring-faithfulness-chain-thought-reasoning/) cuts a model's written reasoning short, inserts mistakes, rewords it, or swaps it for meaningless dots, and then checks whether the final answer shifts. Each intervention isolates one possible failure. These are post-hoc justification, concealed additional processing, and messages encoded in the phrasing itself. Read alongside the opinions study, it shows a turn toward asking what a model's outputs reveal about the model itself. Bylines on these papers list the University of California, Berkeley, the University of Oxford and the Machine Intelligence Research Institute, all between 2022 and 2023. Frequent collaborators include [Esin Durmus](https://md.research.elamur.ai/free/en/authors/esin-durmus/), [Anna Chen](https://md.research.elamur.ai/free/en/authors/anna-chen/) and [Danny Hernandez](https://md.research.elamur.ai/free/en/authors/danny-hernandez/), each on two papers, along with [Ethan Perez](https://md.research.elamur.ai/free/en/authors/ethan-perez/). 3 papers on this site · 2022–2023 current affiliation [Johnson Center for Child Health and Development](https://ror.org/034srtc18), Austin, United States Karina Nguyen has 3 papers on this site, in 2022–2023, most of them on [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/), [Evaluation & benchmarks](https://md.research.elamur.ai/free/en/topics/evaluation-benchmarks/), [Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/); often with [Anna Chen](https://md.research.elamur.ai/free/en/authors/anna-chen/), [Danny Hernandez](https://md.research.elamur.ai/free/en/authors/danny-hernandez/), [Esin Durmus](https://md.research.elamur.ai/free/en/authors/esin-durmus/). [OpenAlex](https://openalex.org/A5032805177) · [**h-index 8**](https://openalex.org/authors/A5032805177) · [14 works](https://openalex.org/works?filter=author.id:A5032805177) · [396 citations](https://openalex.org/works?filter=author.id:A5032805177&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [3 works on this site](https://md.research.elamur.ai/free/en/authors/karina-nguyen/#papers) · [873 citations in our library](https://md.research.elamur.ai/free/en/authors/karina-nguyen/cited-by/) · 1 more work in our library 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=Karina+Nguyen) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Karina+Nguyen) · [dblp](https://dblp.org/search?q=Karina+Nguyen) · [ACL Anthology](https://aclanthology.org/search/?q=Karina+Nguyen) ## Affiliations - [Johnson Center for Child Health and Development](https://ror.org/034srtc18) · Austin, United States —current · 2026 on 2 papers - [Machine Intelligence Research Institute](https://ror.org/01h4ass90) · Berkeley, United States —2022–2023 on 2 papers - [University of California, Berkeley](https://md.research.elamur.ai/free/en/universities/university-of-california-berkeley/) · Berkeley, United States —2023 on 2 papers - [University of Oxford](https://md.research.elamur.ai/free/en/universities/university-of-oxford/) · Oxford, United Kingdom —2023 on 2 papers From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 1 more institution on single bylines are not listed. ## Topics [Language models & pretraining 2](https://md.research.elamur.ai/free/en/topics/language-models/) · [Evaluation & benchmarks 1](https://md.research.elamur.ai/free/en/topics/evaluation-benchmarks/) · [Reasoning 1](https://md.research.elamur.ai/free/en/topics/reasoning/) ## Works by year ### 2020s [**2023** 2 papers](https://md.research.elamur.ai/free/en/authors/karina-nguyen/year/2023/) · [**2022** 1 paper](https://md.research.elamur.ai/free/en/papers/perez-2022-discovering-language-model-behaviors-model/) ## Papers The most cited first (OpenAlex and our library). 1. **[Discovering Language Model Behaviors with Model-Written Evaluations](https://md.research.elamur.ai/free/en/papers/perez-2022-discovering-language-model-behaviors-model/)** *(2022 · 305 citations)* Testing how language models behave usually relies on datasets written by crowdworkers or built from existing sources. 2. **[Towards Measuring the Representation of Subjective Global Opinions in Language Models](https://md.research.elamur.ai/free/en/papers/durmus-2023-towards-measuring-representation-subjective-global/)** *(2023 · 262 citations)* LLMs increasingly answer contested social questions, but it is unclear whose views those answers reflect. The Anthropic authors propose a way to measure this. They built GlobalOpinionQA, a multiple-choice dataset drawn from two cross-national surveys, Pew Global Attitudes and the World Values Survey. Each question… 3. **[Measuring Faithfulness in Chain-of-Thought Reasoning](https://md.research.elamur.ai/free/en/papers/lanham-2023-measuring-faithfulness-chain-thought-reasoning/)** *(2023 · 249 citations)* Step-by-step "Chain-of-Thought" (CoT) prompting makes LLMs more accurate. It is less clear whether the written reasoning actually explains how the model reaches its answer. The authors, mostly at Anthropic, built a set of interventions to test this. They cut the CoT short, inserted an error into it, reworded it, or… ## 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 20 co-authors with a page, the most papers together first. - [Anna Chen](https://md.research.elamur.ai/free/en/authors/anna-chen/) ×2 - [Danny Hernandez](https://md.research.elamur.ai/free/en/authors/danny-hernandez/) ×2 - [Esin Durmus](https://md.research.elamur.ai/free/en/authors/esin-durmus/) ×2 - [Amanda Askell](https://md.research.elamur.ai/free/en/authors/amanda-askell/) - [Andy Jones](https://md.research.elamur.ai/free/en/authors/andy-jones/) - [Anton Bakhtin](https://md.research.elamur.ai/free/en/authors/anton-bakhtin/) - [Carol Chen](https://md.research.elamur.ai/free/en/authors/carol-chen/) - [Catherine Olsson](https://md.research.elamur.ai/free/en/authors/catherine-olsson/) - [Dustin Li](https://md.research.elamur.ai/free/en/authors/dustin-li/) - [Ethan Perez](https://md.research.elamur.ai/free/en/authors/ethan-perez/) - [Jackson Kernion](https://md.research.elamur.ai/free/en/authors/jackson-kernion/) - [Kamile Lukošiute](https://md.research.elamur.ai/free/en/authors/kamile-luko-vsiute/) ## Cited by Papers of our library cite their works 873 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/karina-nguyen/cited-by/).