Spaces:
Running
Running
Update README.md
Browse files
README.md
CHANGED
|
@@ -9,6 +9,7 @@ pinned: false
|
|
| 9 |
<p align="center">
|
| 10 |
<img src="OpenDataArena.PNG" alt="OpenDataArena Banner" width="300">
|
| 11 |
</p>
|
|
|
|
| 12 |
## π About OpenDataArena
|
| 13 |
|
| 14 |
**OpenDataArena (ODA)** is an open research initiative devoted to evaluating, benchmarking, and creating high-value datasets for the post-training era of large language models (LLMs).
|
|
@@ -19,9 +20,9 @@ To make **data evaluation scientific, transparent, and community-driven**, while
|
|
| 19 |
|
| 20 |
### π Key Features
|
| 21 |
|
| 22 |
-
- π **Dataset Leaderboard** β [Leaderboard](https://opendataarena.github.io/leaderboard.html) ranks
|
| 23 |
-
- π **Comprehensive Scoring System** β [Scoring tool](https://github.com/OpenDataArena/OpenDataArena-Tool/tree/main/data_scorer) measures dataset quality, diversity,
|
| 24 |
-
- π§° **Open-Source Toolkit** β *[OpenDataArena-Tool](https://github.com/OpenDataArena/OpenDataArena-Tool)* enables dataset
|
| 25 |
- π± **High-Value Data Generation** β beyond evaluation, ODA continuously produces and shares new, top-quality datasets for fine-tuning and alignment research.
|
| 26 |
|
| 27 |
|
|
|
|
| 9 |
<p align="center">
|
| 10 |
<img src="OpenDataArena.PNG" alt="OpenDataArena Banner" width="300">
|
| 11 |
</p>
|
| 12 |
+
|
| 13 |
## π About OpenDataArena
|
| 14 |
|
| 15 |
**OpenDataArena (ODA)** is an open research initiative devoted to evaluating, benchmarking, and creating high-value datasets for the post-training era of large language models (LLMs).
|
|
|
|
| 20 |
|
| 21 |
### π Key Features
|
| 22 |
|
| 23 |
+
- π **Dataset Leaderboard** β [Leaderboard](https://opendataarena.github.io/leaderboard.html) ranks the most valuable datasets across multiple domains, based on unified post-training benchmarks.
|
| 24 |
+
- π **Comprehensive Scoring System** β [Scoring tool](https://github.com/OpenDataArena/OpenDataArena-Tool/tree/main/data_scorer) measures dataset quality, diversity, and learning values using reproducible pipelines.
|
| 25 |
+
- π§° **Open-Source Toolkit** β *[OpenDataArena-Tool](https://github.com/OpenDataArena/OpenDataArena-Tool)* enables dataset evaluation, scoring with a standardized, community-driven workflow.
|
| 26 |
- π± **High-Value Data Generation** β beyond evaluation, ODA continuously produces and shares new, top-quality datasets for fine-tuning and alignment research.
|
| 27 |
|
| 28 |
|