---
license: cc-by-nc-sa-4.0
dataset_info:
- config_name: corpus
  features:
  - name: doc_id
    dtype: string
  - name: contents
    dtype: string
  - name: metadata
    struct:
    - name: creation_datetime
      dtype: string
    - name: file_name
      dtype: string
    - name: file_path
      dtype: string
    - name: file_size
      dtype: int64
    - name: file_type
      dtype: 'null'
    - name: last_accessed_datetime
      dtype: string
    - name: last_modified_datetime
      dtype: string
  splits:
  - name: train
    num_bytes: 11237324
    num_examples: 8576
  download_size: 4114384
  dataset_size: 11237324
- config_name: qa
  features:
  - name: qid
    dtype: string
  - name: query
    dtype: string
  - name: retrieval_gt
    sequence:
      sequence: string
  - name: generation_gt
    sequence: string
  splits:
  - name: train
    num_bytes: 186908
    num_examples: 520
  download_size: 121089
  dataset_size: 186908
configs:
- config_name: qa
  data_files:
  - split: train
    path: qa/train-*
- config_name: corpus
  data_files:
  - split: train
    path: corpus/train-*

---

## AutoRAG evaluation dataset 

### Made with 2024 LLM resesarch articles (papers)

This dataset is an example for [AutoRAG](https://github.com/Marker-Inc-Korea/AutoRAG).
You can directly use this dataset for optimizng and benchmarking your RAG setup in AutoRAG.

### How this dataset created?

This dataset is 100% synthetically generated by GPT-4 and `Marker Inc.` technology.

At first, we collected 110 latest LLM papers at arxiv. 
We used `Marker` OCR model to extract texts.
And chunk it using MarkdownSplitter and TokenSplitter from Langchain.
For more quality, we delete all `References` in the research articles.
And then, it randomly select 520 passages from chunked corpus for generating question.
At last, our custom pipeline generates various and unique questions with GPT-4. 


## Acknowledgements

This dataset's corpus is originated various LLM related research articles on arixv.
Marker Inc. do not have copyright or any rights about corpus content itself.

Plus, this is a alpha version of our evaluation data generation pipeline without human verification, so its quality might be lower than human-generated dataset.