metadata

license: mit
task_categories:
  - text-classification
language:
  - en

arXiv-model2vec Datase

The arXiv-model2vec dataset contains embeddings for all arXiv paper abstracts and their corresponding titles, generated using the Model2Vec proton-8M model. This dataset is released to support research projects that require high-quality semantic representations of scientific papers.

Dataset Details

Source: The dataset includes embeddings derived from arXiv paper titles and abstracts.
Embedding Model: The embeddings are generated using the Model2Vec proton-8M model, which provides dense vector representations of text.
Content:
- Title Embeddings: Vector representations of the titles of arXiv papers.
- Abstract Embeddings: Vector representations of the abstracts of arXiv papers.

Intended Use Cases

This dataset can be used for a variety of research applications, including:

Semantic Search: Improve search systems by leveraging the semantic information captured in the embeddings.
Clustering and Classification: Group or categorize papers based on their content similarity.
Recommendation Systems: Build systems that recommend relevant papers to users based on their interests.
Trend Analysis: Analyze trends in scientific research by examining changes in the embeddings over time.

Dataset Release

The arXiv-model2vec dataset is made available for academic and research purposes. We hope this resource will aid researchers in exploring new insights and advancing the field of scientific literature analysis.

For any questions or further information, feel free to reach out.