README.md

---
task_categories:
- text-ranking
pretty_name: HPLT2-JQL-Education
size_categories:
- n>1T
language:
- sq
- bg
- ca
- cs
- da
- de
- es
- et
- el
- eu
- fi
- fr
- gl
- ga
- hr
- hu
- hy
- is
- it
- lv
- lt
- mk
- nl
- pl
- pt
- ro
- sl
- sk
- sr
- tr
- sv
- nb
- nn
configs:
- config_name: als_Latn
  data_files:
  - split: train
    path: als_Latn/*
- config_name: bul_Cyrl
  data_files:
  - split: train
    path: bul_Cyrl/*
- config_name: cat_Latn
  data_files:
  - split: train
    path: cat_Latn/*
- config_name: ces_Latn
  data_files:
  - split: train
    path: ces_Latn/*
- config_name: dan_Latn
  data_files:
  - split: train
    path: dan_Latn/*
- config_name: deu_Latn
  data_files:
  - split: train
    path: deu_Latn/*
- config_name: est_Latn
  data_files:
  - split: train
    path: est_Latn/*
- config_name: ell_Grek
  data_files:
  - split: train
    path: ell_Grek/*
- config_name: eus_Latn
  data_files:
  - split: train
    path: eus_Latn/*
- config_name: fin_Latn
  data_files:
  - split: train
    path: fin_Latn/*
- config_name: fra_Latn
  data_files:
  - split: train
    path: fra_Latn/*
- config_name: gle_Latn
  data_files:
  - split: train
    path: gle_Latn/*
- config_name: glg_Latn
  data_files:
  - split: train
    path: glg_Latn/*
- config_name: hrv_Latn
  data_files:
  - split: train
    path: hrv_Latn/*
- config_name: hun_Latn
  data_files:
  - split: train
    path: hun_Latn/*
- config_name: hye_Armn
  data_files:
  - split: train
    path: hye_Armn/*
- config_name: isl_Latn
  data_files:
  - split: train
    path: isl_Latn/*
- config_name: ita_Latn
  data_files:
  - split: train
    path: ita_Latn/*
- config_name: lit_Latn
  data_files:
  - split: train
    path: lit_Latn/*
- config_name: lvs_Latn
  data_files:
  - split: train
    path: lvs_Latn/*
- config_name: mkd_Cyrl
  data_files:
  - split: train
    path: mkd_Cyrl/*
- config_name: nld_Latn
  data_files:
  - split: train
    path: nld_Latn/*
- config_name: nno_Latn
  data_files:
  - split: train
    path: nno_Latn/*
- config_name: nob_Latn
  data_files:
  - split: train
    path: nob_Latn/*
- config_name: pol_Latn
  data_files:
  - split: train
    path: pol_Latn/*
- config_name: por_Latn
  data_files:
  - split: train
    path: por_Latn/*
- config_name: ron_Latn
  data_files:
  - split: train
    path: ron_Latn/*
- config_name: slk_Latn
  data_files:
  - split: train
    path: slk_Latn/*
- config_name: slv_Latn
  data_files:
  - split: train
    path: slv_Latn/*
- config_name: spa_Latn
  data_files:
  - split: train
    path: spa_Latn/*
- config_name: srp_Cyrl
  data_files:
  - split: train
    path: srp_Cyrl/*
- config_name: swe_Latn
  data_files:
  - split: train
    path: swe_Latn/*
- config_name: tur_Latn
  data_files:
  - split: train
    path: tur_Latn/*
- config_name: ukr_Cyrl
  data_files:
  - split: train
    path: ukr_Cyrl/*
---

# HPLT2-Edu-scores

## Dataset summary

HPLT2-JQL-Education is a **model-annotated** language subset of [**HPLT2**](https://hplt-project.org/datasets/v2.0), spanning **35 languages**. 
Our model-annotations allow for a filtering that achieves higher-quality training outcomes without excessively aggressive data reduction. 
The original FW2 heuristic filtering method serves as our baseline, providing reference points for both the volume of retained tokens and downstream model performance.
For example, in the Spanish language case, applying the 0.6 threshold retains over 9% more tokens than FW2 train while still surpassing its quality . 

HPLT2-Edu-scores was created based on scores assigned by a deep learning classifier trained to identify **educational samples** using [**Snowflake's Arctic-embed-m-v2.0**](https://huggingface.co/Snowflake/snowflake-arctic-embed-m-v2.0) embeddings.

For all training ablations, we used dense decoder-only models with **2 billion parameters**, following the LLaMA architecture.
For more details, see our paper [https://arxiv.org/abs/2505.22232](https://arxiv.org/abs/2505.22232).

The approach as described in the paper is easy to extend to other languages as well, and we might consider adding new languages to an upcoming version of the present dataset.

We also separately release the computed general-purpose embedding vectors for the the full sets of the original HPLT2 dataset, in the respective languages, as they can be useful for other applications beyond quality filtering: [HPLT2-embeddings](https://huggingface.co/datasets/JQL-AI/hplt2_embeddings).

# Dataset Structure

## Data Fields
Each data entry includes:
- `score_Gemma_Snowflake`: Quality score obtained by the Gemma-based Snowflake classifier
- `score_Llama_Snowflake`: Quality score obtained by the Llama-based Snowflake classifier
- `score_Mistral_Snowflake`: Quality score obtained by the Mistral-based Snowflake classifier
- `source_filename`: Source filename of the original file.


## Data Instance
```json
{
  "id": "0",
  "file_path": "/leonardo_scratch/large/userexternal/mfromm00/data/raw_data/HPLT2/output/embeddings/als_Latn/als_Latn/000_000_00000.jsonl.h5",
  "document_id": "29d82196d55803ab9c792e45b59919bf_0",
  "source_filename": "als_Latn/als_Latn/000_000_00000.jsonl.h5",
  "score_Gemma_Snowflake": 0.330078125,
  "score_Llama_Snowflake": -0.34765625,
  "score_Mistral_Snowflake": -0.390625
}
```


## Origin of the Dataset

This dataset, derived from HPLT2, includes web content collected from 2012 to 2023. As HPLT2 is sourced from the broader internet, it may contain some personally identifiable information (PII), despite efforts to anonymize email addresses and public IP addresses during processing.

## Considerations for Data Usage

For information on social impact, potential biases, and known limitations, please refer to the  [HPLT2 documentation](https://hplt-project.org/datasets/v2.0).


## Citation information
If you use this dataset in your research or applications, please use the following citation:
```
@article{ali2025judging,
    title     = {Judging Quality Across Languages: A Multilingual Approach to Pretraining Data Filtering with Language Models},
    author    = {
      Mehdi Ali,
      Manuel Brack,
      Max Lübbering,
      Elias Wendt,
      Abbas Goher Khan,
      Richard Rutmann,
      Alex Jude,
      Maurice Kraus,
      Alexander Arno Weber,
      Felix Stollenwerk,
      David Kaczér,
      Florian Mai,
      Lucie Flek,
      Rafet Sifa,
      Nicolas Flores-Herr,
      Joachim Köhler,
      Patrick Schramowski,
      Michael Fromm,
      Kristian Kersting
    },
    year      = {2025},
    journal   = {arXiv preprint arXiv:2505:22232}
  }

```