Get the app
extract gated 0 0

lvogel123/m2s-mhj

published by @lvogel123

Alignment dataset tracked in the /datasets sub-catalog.

Lifetime downloads
5K
Last 30 days
747
Likes
-
Size
1.4 MB
Created on HF
2025-10-05
Age
1 years ago

Description

MHJ → M2S (single-turn)

License: CC-BY-NC-4.0.Sources: Derived from the Multi-Turn Human Jailbreaks (MHJ) dataset and transformed using the M2S procedure (“Hyphenize / Numberize / Pythonize”).

We convert MHJ multi-turn jailbreak conversations into single-turn prompts following Ha et al., 2025. This release includes only IDs and prompt text (no model outputs/labels). Use is for research and non-commercial purposes only; redistribution must retain this license and attributions.… See the full description on the dataset page: https://huggingface.co/datasets/lvogel123/m2s-mhj.

Tags

task_categories:text-generationsource_datasets:ScaleAI/mhjlanguage:enlicense:cc-by-nc-4.0size_categories:1K<n<10Kformat:parquetmodality:textlibrary:datasetslibrary:pandaslibrary:mlcroissantlibrary:polarsregion:usjailbreaksafetyadversarial-robustness

README current version from Hugging Face

pretty_name: MHJ → M2S (single-turn)
license: cc-by-nc-4.0
language: [en]
tags: [jailbreak, safety, adversarial-robustness]
task_categories: [text-generation]
source_datasets:

  • ScaleAI/mhj

MHJ → M2S (single-turn)

License: CC-BY-NC-4.0.
Sources: Derived from the Multi-Turn Human Jailbreaks (MHJ) dataset and transformed using the M2S procedure (“Hyphenize / Numberize / Pythonize”).

We convert MHJ multi-turn jailbreak conversations into single-turn prompts following Ha et al., 2025. This release includes only IDs and prompt text (no model outputs/labels). Use is for research and non-commercial purposes only; redistribution must retain this license and attributions.

Attribution

  • MHJ dataset (primary source): Scale AI, Multi-Turn Human Jailbreaks (MHJ). CC-BY-NC-4.0.
  • M2S method/repo: Ha et al., One-Shot is Enough: Consolidating Multi-Turn Attacks into Efficient Single-Turn Prompts for LLMs (ACL 2025) and associated GitHub repo.

Modifications

  • Transformed dialogues → single-turn prompts; removed upstream model outputs/scores.

Ethical & Safety Notes

Contains adversarial prompts for robustness testing

README history 5 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. 2025-10-05Create LICENSEe6fc9b51.2 KB
    Loading...
  2. 2025-10-05Update README.md00b792a1.2 KB
    Loading...
  3. 2025-10-05Update README.mdb6d6113513 B
    Loading...
  4. 2025-10-05Upload dataset33f4e51580 B
    Loading...
  5. 2025-10-05initial commit5c7d3810 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. 2025-10-06[bot] Conversion to Parquetopen1 💬#1
    Loading...

Continue on Hugging Face →

Related datasets same stage · extract

Anthropic/hh-rlhf
2.1M dl 2442 likes 90.4 MB
Dataset Card for HH-RLHF Dataset Summary This repository provides access to two different kinds of data: Human ...
djain95/sae-jailbreaks-cache
468.6K dl 1 likes 30.6 GB
mosaicml/dolly_hhrlhf
309.5K dl 113 likes 23.7 MB
Dataset Card for "dolly_hhrlhf" This dataset is a combination of Databrick's dolly-15k dataset and a filtered subset of Anthrop...
Dahoas/full-hh-rlhf
266.3K dl 104 likes 129.8 MB
Dataset Card for "full-hh-rlhf" Anthropic's HH dataset reformatted into prompt, chosen, rejected samples.
mlabonne/harmful_behaviors
198.2K dl 169 likes 23 KB
mlabonne/harmless_alpaca
190.2K dl 51 likes 1.2 MB
walledai/AdvBench
176.3K dl 123 likes 38 KB
Dataset Card for AdvBench Paper: Universal and Transferable Adversarial Attacks on Aligned Language Models Data: AdvBench Datas...
JailbreakV-28K/JailBreakV-28k
116.4K dl 80 likes 288.8 MB
⛓‍💥 JailBreakV-28K: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks 🌐 ...

Read further

Community best practices (extraction)
5 rules the field converged on for building the pair.
Building your own extraction pair
When the reference pair is not enough. Sizing, framing, and evaluation.
Datasets that shaped abliteration
The five corpora that did most of the work.
Non-English abliteration communities
How Chinese, Russian, Japanese, and Korean practitioners build extraction pairs.

Metadata

License
cc-by-nc-4.0

Repository

DIR data 3 files FILE .gitattributes FILE LICENSE FILE README.md

File contents live on Hugging Face. Click a file or folder to open it there.

Download the dataset

Gated dataset. You need to accept the author's terms on Hugging Face before downloading. Visit the repository, click Agree and access repository, then log in locally with huggingface-cli login and use one of the snippets below.

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("lvogel123/m2s-mhj")
Raw snapshot (Python) 
from huggingface_hub import snapshot_download
snapshot_download(repo_id="lvogel123/m2s-mhj", repo_type="dataset")
git clone (requires git-lfs) 
git clone https://huggingface.co/datasets/lvogel123/m2s-mhj
Browse repository on Hugging Face ->
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.

Get the app Browse the catalog
Open in Abliteration