Summary of Latxa: An Open Language Model and Evaluation Suite For Basque, by Julen Etxaniz et al.
Latxa: An Open Language Model and Evaluation Suite for Basque
by Julen Etxaniz, Oscar Sainz, Naiara Perez, Itziar Aldabe, German Rigau, Eneko Agirre, Aitor Ormazabal, Mikel Artetxe, Aitor Soroa
First submitted to arxiv on: 29 Mar 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel family of large language models, Latxa, is introduced for the Basque language, ranging from 7 billion to 70 billion parameters. This paper builds upon Llama 2 and pretrains it on a new Basque corpus with 4.3 million documents and 4.2 billion tokens. To address the scarcity of high-quality benchmarks for Basque, four multiple-choice evaluation datasets are introduced: EusProficiency, EusReading, EusTrivia, and EusExams. Latxa outperforms previous open models by a large margin in an extensive evaluation. It is competitive with GPT-4 Turbo in language proficiency and understanding, despite lagging behind in reading comprehension and knowledge-intensive tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Latxa is a new type of computer model that can understand and generate text in the Basque language. The researchers made this model by training it on a large collection of texts written in Basque. They also created four new tests to see how well Latxa could do, compared to other models that were already available. These tests were designed to test different skills, such as understanding and generating text. Latxa did very well on the tests, especially for tasks like language proficiency and understanding. |
Keywords
» Artificial intelligence » Gpt » Llama
