Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -482,7 +482,7 @@ fw = load_dataset("HuggingFaceFW/fineweb-edu", name="CC-MAIN-2024-10", split="tr
|
|
482 |
```
|
483 |
|
484 |
## Dataset curation
|
485 |
-
A new approach has recently emerged for filtering LLM training datasets: using synthetic data to develop classifiers for identifying educational content. This technique was used in the trainings of [LLama3](https://ai.meta.com/blog/meta-llama-3-meta-ai-responsibility/)
|
486 |
|
487 |
The highly popular Phi3 models were trained on 3.3 and 4.8 trillion tokens, with the paper stating: “Our training data consists of heavily filtered publicly available web data (according to the 'educational level') from various open internet sources, as well as synthetic LLM-generated data". Similarly, the LLama3 blog post notes: “We found that previous generations of Llama are good at identifying high-quality data, so we used Llama 2 to help build the text-quality classifiers that are powering Llama 3.” However these classifiers and filtered datasets are not publicly available. To enhance FineWeb's quality, we developed an educational quality classifier using annotations generated by [LLama3-70B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) to create FineWeb-Edu.
|
488 |
|
|
|
482 |
```
|
483 |
|
484 |
## Dataset curation
|
485 |
+
A new approach has recently emerged for filtering LLM training datasets: using synthetic data to develop classifiers for identifying educational content. This technique was used in the trainings of [LLama3](https://ai.meta.com/blog/meta-llama-3-meta-ai-responsibility/) and [Phi3](https://arxiv.org/abs/2404.14219), but its large-scale impact on web data filtering hasn't been fully explored or published.
|
486 |
|
487 |
The highly popular Phi3 models were trained on 3.3 and 4.8 trillion tokens, with the paper stating: “Our training data consists of heavily filtered publicly available web data (according to the 'educational level') from various open internet sources, as well as synthetic LLM-generated data". Similarly, the LLama3 blog post notes: “We found that previous generations of Llama are good at identifying high-quality data, so we used Llama 2 to help build the text-quality classifiers that are powering Llama 3.” However these classifiers and filtered datasets are not publicly available. To enhance FineWeb's quality, we developed an educational quality classifier using annotations generated by [LLama3-70B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) to create FineWeb-Edu.
|
488 |
|