---
title: "Argentina — research institutions, scientists and papers"
description: "Argentina: 2 institutions with papers in our library in 1 agglomerations, 0 people with a page, 3 works, 828 citations."
canonical: "https://research.elamur.ai/free/en/countries/argentina/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/countries/argentina/)

[home](https://md.research.elamur.ai/free/en/) / [countries](https://md.research.elamur.ai/free/en/countries/) / Argentina

# Argentina

Country 2 institutions in 1 agglomeration · [all countries](https://md.research.elamur.ai/free/en/countries/)

[**2** institutions with a page](https://md.research.elamur.ai/free/en/countries/argentina/#institutions) · [**1** agglomeration](https://md.research.elamur.ai/free/en/countries/argentina/#agglomerations) · [**0** people with a page](https://md.research.elamur.ai/free/en/countries/argentina/#authors) · [**3** works in our library](https://md.research.elamur.ai/free/en/countries/argentina/#works) · [**828** citations of their works](https://md.research.elamur.ai/free/en/countries/argentina/#works)

## About

### Profile

Argentina has a small but focused presence on this site. Its research here comes from institutions in a single urban region, [Buenos Aires](https://md.research.elamur.ai/free/en/agglomerations/buenos-aires/). The country is represented by a university, the [Universidad de Buenos Aires](https://md.research.elamur.ai/free/en/universities/universidad-de-buenos-aires/), and a government research body, the [Consejo Nacional de Investigaciones Científicas y Técnicas](https://md.research.elamur.ai/free/en/government/consejo-nacional-de-investigaciones-cientificas-y-tecnicas/). The university appears on every Argentine work in the library. The national council shares authorship on the two most cited of them.

Among the works here, Argentine research centres on language and computation. It covers how large language models are tested, how that testing can be fair across languages and cultures, and how the meaning of language can be recovered from brain activity. The library groups these works under [natural language processing](https://md.research.elamur.ai/free/en/shelves/nlp/), [linguistics and languages](https://md.research.elamur.ai/free/en/shelves/linguistics/), and [mathematics and statistics](https://md.research.elamur.ai/free/en/shelves/mathematics/). Readers who want to compare Argentina with other nations can browse the full list of [countries](https://md.research.elamur.ai/free/en/countries/). The country's institutions are gathered on a [dedicated index](https://md.research.elamur.ai/free/en/countries/argentina/institutions/), and its urban regions on a [separate one](https://md.research.elamur.ai/free/en/countries/argentina/agglomerations/).

### History

Argentina's record on this site is recent. Its earliest works here date from 2025, and the most recent from 2026. In that short span the research follows two related lines.

The first line, in 2025, belongs to large international efforts to evaluate language models. Both 2025 works ask how well current benchmarks measure what models can actually do. One asks the question in terms of difficulty and the other in terms of linguistic and cultural coverage. The second line appears in 2026 and moves toward cognitive neuroscience. It concerns the decoding of language from functional brain imaging, with language models used as part of the method.

The library shows no earlier Argentine works, so it cannot describe how research in the country developed before 2025. What it does show is a quick turn from contributing to large evaluation projects toward a study in which language models serve as instruments for neuroscience.

### Research profile

The clearest strength among the works here is the evaluation of large language models. [Humanity's Last Exam](https://md.research.elamur.ai/free/en/papers/phan-2025-humanity-s-last-exam/) starts from the problem that well-known benchmarks such as MMLU have become nearly saturated. Once the strongest systems score close to the ceiling on such tests, the tests reveal little about real capability. The project responds by building a much harder body of questions intended to separate the abilities of advanced models more clearly. It is the most cited Argentine work in the library, and other papers in our library draw on it often.

A closely related line deals with fairness and coverage across languages. [Global MMLU](https://md.research.elamur.ai/free/en/papers/singh-2025-global-mmlu-understanding-addressing-cultural/) examines cultural and linguistic bias in multilingual evaluation. Its authors measure how far existing evaluation material carries assumptions tied to particular cultures and languages. They then build an alternative designed to give a more balanced view of model performance across many languages. This work is why the library places Argentina under the topic of [multilingual and cross-lingual NLP](https://md.research.elamur.ai/free/en/topics/multilingual-nlp/). It also links Argentine research to the wider study of linguistics and languages.

The third line joins language technology with brain science. [Topic-context Dependency on Continuous Semantic Reconstruction of Language from fMRI Signals](https://md.research.elamur.ai/free/en/papers/travi-2026-topic-context-dependency-continuous-semantic/) studies brain-to-text decoders. In these systems a language model proposes possible continuations of a text, and a voxelwise encoding model predicts the fMRI response each candidate would produce. The study looks at how the topic and context of the language affect the quality of this semantic reconstruction. Its use of statistical encoding models and signal analysis helps explain why the library also places Argentine work under mathematics and statistics.

Together these works point to a research profile organised around language. It covers language as a test of machine capability, language in its full cultural and linguistic diversity, and language as something the brain encodes and that might be read back from imaging. On this site the Argentine contribution does not extend into fields outside this cluster.

### Leading scientists

The scientists first associated with Argentine institutions on this site are linked through the Universidad de Buenos Aires.

[Horacio Saggion](https://md.research.elamur.ai/free/en/authors/horacio-saggion/) is the first scientist connected with the university. Given the institution's works here, Saggion's place falls within the language-centred research that defines the Argentine profile. That research is concerned with natural language processing and the evaluation of language technology across languages.

[Melanie Sclar](https://md.research.elamur.ai/free/en/authors/melanie-sclar/) is connected both with the Universidad de Buenos Aires and with the Consejo Nacional de Investigaciones Científicas y Técnicas. Sclar is the one figure on this site who links the university and the national council. Those two institutions share the 2025 works on benchmarking and multilingual evaluation, so Sclar's connection lies along that line of language-model assessment.

Beyond these two, the bylines of the Argentine works name many other researchers. Humanity's Last Exam is led by Long Phan, with Alice Gatti and Ziwen Han among its early authors. Global MMLU is led by Shivalika Singh, with Angelika Romanou and Clémentine Fourrier among the first names. The fMRI study is led by Fermin Travi, with Agustin Delmagro and Diego Fernandez Slezak among its authors. The library does not say which of these co-authors work at the Argentine institutions themselves.

### Landmark works

The most cited Argentine work in our library is Humanity's Last Exam, from 2025. Its importance lies in the problem it tackles: once leading benchmarks are saturated, the field loses a reliable way to compare the newest models. By proposing a far more demanding test, the work offers a fresh reference point. Other papers in the library cite it widely, which makes it the main point of contact between Argentine research and the broader literature here.

Global MMLU, also from 2025, is the second most cited. It moves the evaluation question from difficulty to fairness. Its contribution is to show that widely used evaluation material is not culturally neutral, and to supply a multilingual alternative that takes those biases into account. It too is cited by many papers in our library, especially within multilingual research.

The 2026 fMRI study is the newest of the Argentine works and has not yet gathered citations in the library. It stands out for applying language models to neuroscience, where they serve as part of the method rather than as the subject of evaluation. Its question about topic and context in brain-based semantic reconstruction opens a direction that the earlier works do not cover.

### Labs and units

On this site Argentina's research is carried by its university and its national council, both based in Buenos Aires. The Universidad de Buenos Aires is the broader of the two and appears on every work, from the benchmark studies to the brain decoding research. The Consejo Nacional de Investigaciones Científicas y Técnicas is a government research body. Its works here are the 2025 evaluation studies, which it shares with the university. The library does not describe any departments, laboratories or smaller units within these institutions.

### Collaborations

Collaboration is central to Argentina's presence on this site. The two most cited works are large multi-author projects. Humanity's Last Exam and Global MMLU each bring together long lists of researchers. Argentine institutions are part of these wider efforts rather than their sole source.

Within the country, the main link is between the Universidad de Buenos Aires and the Consejo Nacional de Investigaciones Científicas y Técnicas. The two share authorship on the evaluation works, and Melanie Sclar is associated with both. The fMRI study appears on this site under the university alone. It draws on a group of authors including Fermin Travi, Agustin Delmagro and Diego Fernandez Slezak. The library does not show company partners for Argentina, so no link between industry and academia can be traced here.

### Timeline

In 2025 Argentine research appears on this site through two large evaluation projects on language models. Humanity's Last Exam introduces a much harder benchmark in response to the saturation of familiar tests. Global MMLU addresses cultural and linguistic bias in multilingual evaluation. Both involve the Universidad de Buenos Aires and the Consejo Nacional de Investigaciones Científicas y Técnicas, and both quickly became widely cited in the library.

In 2026 the Argentine record widens toward cognitive neuroscience. The study by Fermin Travi and colleagues examines how topic and context shape the reconstruction of meaning from fMRI signals, using language models inside a decoding system. It appears under the Universidad de Buenos Aires. It shows the same interest in language seen in the earlier work, now applied to the study of the brain.

## Where

## Agglomerations (1)

1. [**Buenos Aires**2 institutions*828 citations*](https://md.research.elamur.ai/free/en/agglomerations/buenos-aires/)

## Institutions (2)

### 🎓 Universities (1)

**[Universidad de Buenos Aires](https://md.research.elamur.ai/free/en/universities/universidad-de-buenos-aires/)**University · [Buenos Aires](https://md.research.elamur.ai/free/en/agglomerations/buenos-aires/), Argentina*3 works · 414 citations*

### 🏛️ Government research bodies (1)

**[Consejo Nacional de Investigaciones Científicas y Técnicas](https://md.research.elamur.ai/free/en/government/consejo-nacional-de-investigaciones-cientificas-y-tecnicas/)**Government body · [Buenos Aires](https://md.research.elamur.ai/free/en/agglomerations/buenos-aires/), Argentina*2 works · 414 citations*

## Works in our library (3)

Works of its institutions, the most cited first.

1. [**Humanity's Last Exam**Long Phan, Alice Gatti, Ziwen Han, et al.*2025 · 235 citations*](https://md.research.elamur.ai/free/en/papers/phan-2025-humanity-s-last-exam/)
2. [**Global MMLU: Understanding and Addressing Cultural and Linguistic Biases in Multilingual Evaluation**Shivalika Singh, Angelika Romanou, Clémentine Fourrier, et al.*2025 · 179 citations*](https://md.research.elamur.ai/free/en/papers/singh-2025-global-mmlu-understanding-addressing-cultural/)
3. [**Topic-context Dependency on Continuous Semantic Reconstruction of Language from fMRI Signals**Fermin Travi, Agustin Delmagro, Diego Fernandez Slezak, et al.*2026*](https://md.research.elamur.ai/free/en/papers/travi-2026-topic-context-dependency-continuous-semantic/)

## Concepts, shelves and topics

Of its works in our library: the number of works on each and its rating among them.

📚 Shelves 3[Mathematics & statistics · 1 work · 15.3](https://md.research.elamur.ai/free/en/shelves/mathematics/) · [Natural language processing · 1 work · 13.4](https://md.research.elamur.ai/free/en/shelves/nlp/) · [Linguistics & languages · 1 work · 0.0](https://md.research.elamur.ai/free/en/shelves/linguistics/)

🧭 Topics 1[Multilingual & cross-lingual NLP · 1 work · 13.4](https://md.research.elamur.ai/free/en/topics/multilingual-nlp/)

## Works by year

### 2020s

[**2026** 1 work](https://md.research.elamur.ai/free/en/papers/travi-2026-topic-context-dependency-continuous-semantic/) · [**2025** 2 works](https://md.research.elamur.ai/free/en/countries/argentina/year/2025/)

- [← back to](https://md.research.elamur.ai/free/en/countries/) — countries