README.md

metadata

license: mit
task_categories:
  - text-retrieval
language:
  - en
size_categories:
  - 1M<n<10M
dataset_info:
  features:
    - name: id
      dtype: string
    - name: url
      dtype: string
    - name: title
      dtype: string
    - name: text
      dtype: string
    - name: edu_quality
      dtype: float64
    - name: naive_quality
      dtype: int64
  splits:
    - name: train
      num_bytes: 1820347708
      num_examples: 1041047
  download_size: 1104502937
  dataset_size: 1820347708
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

MTR Document Collection

1,041,047 Wikipedia passages used as the retrieval corpus for the MTR benchmark.

Part of MTR-Suite (ACL 2026 Main).

Related Resources

Resource	Link	Description
📊 MTR Benchmark	OkayestProgrammer/MTR-BENCH	Original test set
🏋️ MTR Training	OkayestProgrammer/MTR-train	Original training set
🆕 12-Turn Dataset	OkayestProgrammer/mtr-qwen35-fp8-12turn	10K 12-turn conversations with topic switches
🔧 Code	OkayestProgrammer/mtr-suite	Full pipeline code

Columns

Column	Type	Description
id	string	Wikipedia article ID
url	string	Source URL
title	string	Article title
text	string	Passage text (max 2048 chars)
edu_quality	float	Educational quality score
naive_quality	int	Naive quality label

Usage

from datasets import load_dataset

docs = load_dataset("OkayestProgrammer/MTR-DOCUMENT", split="train")
print(f"{len(docs)} documents")
print(docs[0]["title"], "-", docs[0]["text"][:100])

How this corpus is used

The ground_truth_document_idx field in the MTR query datasets (e.g., MTR-BENCH, mtr-qwen35-fp8-12turn) is a row index into this document collection. During evaluation:

1. Embed all 1,041,047 documents → build FAISS index
2. Embed test queries → search the index
3. Compare retrieved document indices against ground_truth_document_idx

See the eval guide for step-by-step instructions.

Citation

@inproceedings{mtr-suite-2026,
    title={MTR-Suite: A Data Synthesis Pipeline, Benchmark, and Models for Conversational Retrieval},
    author={},
    booktitle={Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026)},
    year={2026}
}