Update README.md

README.md

@@ -3,5 +3,55 @@ language:
 - en
 ---
 
-## Arxiv Dataset used by MTEB Arena
-TODO
+# mteb/arena-arxiv-7-2-24 Dataset
+
+## Overview
+
+This dataset, `mteb/arena-arxiv-7-2-24`, is a comprehensive collection of scientific papers from ArXiv up to July 2, 2024. It is designed for use in the MTEB (Massive Text Embedding Benchmark) arena, where various embedding models compete and are ranked based on their performance.
+
+## What is ArXiv?
+
+ArXiv (pronounced "archive") is a free distribution service and open-access archive for scholarly articles. Founded in 1991, it has become a crucial platform for researchers to share their work quickly and efficiently, particularly in the fields of physics, mathematics, computer science, quantitative biology, quantitative finance, statistics, electrical engineering and systems science, and economics.
+
+ArXiv allows researchers to upload preprints (versions of academic papers before peer review) as well as post-prints (versions after peer review). This enables rapid dissemination of new research findings and fosters open scientific communication.
+
+## Dataset Structure
+
+Each instance in the dataset represents a single paper from ArXiv and contains the following fields:
+
+1. **id** (string): A unique identifier for the paper, typically in the format "YYMM.NNNNN" where YY is the year, MM is the month, and NNNNN is a sequential number.
+
+2. **title** (string): The full title of the paper.
+
+3. **abstract** (string): A summary of the paper's content, typically written by the authors.
+
+4. **categories** (string): The ArXiv categories associated with the paper. Papers can belong to multiple categories, reflecting their interdisciplinary nature.
+
+## Example Instance
+
+Here's an example of what a single instance in the dataset looks like:
+
+```json
+{
+  "id": "0704.0001",
+  "title": "Calculation of prompt diphoton production cross sections at Tevatron and LHC energies",
+  "abstract": "A fully differential calculation in perturbative quantum chromodynamics is presented for the production of massive photon pairs at hadron colliders. All next-to-leading order perturbative contributions from quark-antiquark, gluon-(anti)quark, and gluon-gluon subprocesses are included, as well as all-orders resummation of initial-state gluon radiation valid at next-to-next-to-leading logarithmic accuracy. The region of phase space is specified in which the calculation is most reliable. Good agreement is demonstrated with data from the Fermilab Tevatron, and predictions are made for more detailed tests with CDF and DO data. Predictions are shown for distributions of diphoton pairs produced at the energy of the Large Hadron Collider (LHC). Distributions of the diphoton pair transverse momentum and rapidity are presented, as well as the rapidity distribution of the diphoton system. The direct photon contribution is shown separately from the photon fragmentation contributions. The fragmentation contributions are approximated by the use of the first order perturbative contribution only.",
+  "categories": "hep-ph"
+}
+```
+
+## Usage
+
+This dataset is primarily intended for training and evaluating embedding models in the MTEB arena. Researchers and developers can use it to:
+
+1. Train new embedding models on a diverse set of scientific texts.
+2. Evaluate the performance of existing embedding models on scientific literature.
+3. Conduct research on topic modeling, document classification, or information retrieval in the scientific domain.
+
+## Ethical Considerations
+
+When using this dataset, please be aware of potential biases in the scientific literature and the limitations of using preprint data. Not all papers in ArXiv have undergone peer review, so the quality and accuracy of the content may vary.
+
+## Updates and Maintenance
+
+This dataset represents ArXiv papers up to July 2, 2024. For instructions on how to create this dataset again with newer data, please refer to the [create_index_chunks.py script](https://github.com/embeddings-benchmark/arena/blob/main/retrieval/create_index_chunks.py#L107) in the embeddings-benchmark/arena repository.