evaluate 0 0 0

VGraf/ifeval_mt

published by @VGraf

Alignment dataset tracked in the /datasets sub-catalog.

Lifetime downloads
1.6K
Last 30 days
328
Likes
-
Size
5.8 MB
Created on HF
2025-06-27
Age
1 years ago

Tags

size_categories:1K<n<10Kformat:parquetmodality:tabularmodality:textlibrary:datasetslibrary:pandaslibrary:mlcroissantlibrary:polarsregion:us

README current version from Hugging Face


dataset_info:

  • config_name: ood_wildchat_unused_withRewrite
    features:
    • name: id
      dtype: int64
    • name: key
      dtype: int64
    • name: prompt
      dtype: string
    • name: instruction_id_list
      sequence: string
    • name: kwargs
      list:
      • name: N
        dtype: int64
      • name: small_n
        dtype: int64
      • name: percentage
        dtype: int64
      • name: word
        dtype: string
      • name: sep
        dtype: string
      • name: n
        dtype: 'null'
      • name: m
        dtype: 'null'
      • name: keyword
        dtype: string
      • name: min_words
        dtype: int64
      • name: max_words
        dtype: int64
      • name: keyword1
        dtype: 'null'
      • name: keyword2
        dtype: 'null'
      • name: keyword3
        dtype: 'null'
      • name: keyword4
        dtype: 'null'
    • name: messages
      list:
      • name: role
        dtype: string
      • name: content
        dtype: string
        splits:
    • name: test
      num_bytes: 6264928
      num_examples: 1387
      download_size: 2690403
      dataset_size: 6264928
  • config_name: wildchat_unused_withRewrite
    features:
    • name: id
      dtype: int64
    • name: key
      dtype: int64
    • name: prompt
      dtype: string
    • name: instruction_id_list
      sequence: string
    • name: kwargs
      list:
      • name: keyword
        dtype: string
      • name: end_phrase
        dtype: string
      • name: keywords
        sequence: string
      • name: forbidden_words
        sequence: string
      • name: language
        dtype: string
      • name: frequency
        dtype: int64
      • name: relation
        dtype: string
      • name: num_sentences
        dtype: int64
      • name: num_paragraphs
        dtype: int64
      • name: let_frequency
        dtype: int64
      • name: letter
        dtype: string
      • name: num_bullets
        dtype: int64
      • name: capital_frequency
        dtype: int64
      • name: capital_relation
        dtype: string
      • name: num_sections
        dtype: int64
      • name: section_spliter
        dtype: string
      • name: num_words
        dtype: int64
      • name: postscript_marker
        dtype: string
      • name: num_highlights
        dtype: int64
      • name: num_placeholders
        dtype: int64
    • name: messages
      list:
      • name: role
        dtype: string
      • name: content
        dtype: string
        splits:
    • name: test
      num_bytes: 8320918
      num_examples: 1774
      download_size: 3377514
      dataset_size: 8320918
      configs:
  • config_name: ood_wildchat_unused_withRewrite
    data_files:
    • split: test
      path: ood_wildchat_unused_withRewrite/test-*
  • config_name: wildchat_unused_withRewrite
    data_files:
    • split: test
      path: wildchat_unused_withRewrite/test-*

README history 6 revisions

Every night we snapshot the README of every dataset in the catalog. When the SHA changes we archive the new version and diff it against the last. This is the evolving thought record of the alignment-data field: what the author decided to say about the corpus, and how that framing shifted over time.

  1. 2025-07-02Update README.mdc0845ba2.6 KB
    Loading...
  2. 2025-07-02Update README.mdb8736cc2.5 KB
    Loading...
  3. 2025-07-02Update README.mdddcc2372.5 KB
    Loading...
  4. 2025-06-27Upload dataset7307f092.6 KB
    Loading...
  5. 2025-06-27Upload dataset67a44511.5 KB
    Loading...
  6. 2025-06-27initial commit552516e0 B
    Loading...

Used by abliterated models none yet

No abliterated models in our catalog have declared this dataset in their YAML metadata yet. This can mean the dataset is used but not declared, is used outside abliteration workflows, or is new. The link will populate automatically as models are indexed.

Community discussions 1 thread · 1 open · 1 comments

Threads posted by the community on Hugging Face: bug reports, questions about usage, pull requests on the README, license clarifications. Click any thread to expand and read the full conversation inline. Newest first.

  1. 2025-06-28[bot] Conversion to Parquetopen1 💬#1
    Loading...

Related datasets same stage · evaluate

Read further

Repository

Download the dataset

0

Copy any of these snippets into your notebook or terminal. All three fetch directly from Hugging Face using your own credentials.

Recommended: datasets library (Python)
from datasets import load_dataset
ds = load_dataset("VGraf/ifeval_mt")
Raw snapshot (Python)
from huggingface_hub import snapshot_download
snapshot_download(repo_id="VGraf/ifeval_mt", repo_type="dataset")
git clone (requires git-lfs)
git clone https://huggingface.co/datasets/VGraf/ifeval_mt
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration