Get the app
extract 0 0 0

LoneResearch/thinking-steering-vectors

published by @LoneResearch

Alignment dataset tracked in the /datasets sub-catalog.

Lifetime downloads
2K
Last 30 days
508
Likes
-
Size
1.4 GB
Created on HF
2025-09-08
Age
1 years ago

Tags

size_categories:1K<n<10Kformat:jsonlibrary:datasetslibrary:dasklibrary:mlcroissantregion:us

README history 20 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. 2025-10-07Upload folder using huggingface_huba99d3300 B
    Loading...
  2. 2025-10-07Upload folder using huggingface_hubf0e04490 B
    Loading...
  3. 2025-10-07Upload folder using huggingface_hub8ce95100 B
    Loading...
  4. 2025-10-07Upload folder using huggingface_hube53d7b60 B
    Loading...
  5. 2025-10-07Upload folder using huggingface_hub5cb0a160 B
    Loading...
  6. 2025-10-07Upload folder using huggingface_hub2f104ec0 B
    Loading...
  7. 2025-10-07Upload folder using huggingface_hub6cbe9f00 B
    Loading...
  8. 2025-10-07Upload folder using huggingface_hub32ac8130 B
    Loading...
  9. 2025-10-07Upload folder using huggingface_hub86f96210 B
    Loading...
  10. 2025-10-07Upload folder using huggingface_hub205f0ce0 B
    Loading...
  11. 2025-10-07Upload folder using huggingface_hub77f21bd0 B
    Loading...
  12. 2025-10-07Upload folder using huggingface_hubf8e7ee00 B
    Loading...
  13. 2025-10-07Upload folder using huggingface_hub7e748760 B
    Loading...
  14. 2025-10-07Upload folder using huggingface_hub781c14b0 B
    Loading...
  15. 2025-10-07Upload folder using huggingface_hubb78125a0 B
    Loading...
  16. 2025-10-07Upload folder using huggingface_hub15d0e1b0 B
    Loading...
  17. 2025-09-23Upload folder using huggingface_hubc4f81ad0 B
    Loading...
  18. 2025-09-23Upload folder using huggingface_hub15cf3d20 B
    Loading...
  19. 2025-09-23Upload folder using huggingface_hub6e1ff310 B
    Loading...
  20. 2025-09-23Upload folder using huggingface_hub718a64c0 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. 2025-09-22[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.

Repository

DIR DeepSeek-R1-Distill-LLama-8B 47 files DIR DeepSeek-R1-Distill-Qwen-7B 42 files DIR Qwen3-4B 32 files DIR Qwen3-8B 44 files FILE .gitattributes

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

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("LoneResearch/thinking-steering-vectors")
Raw snapshot (Python) 
from huggingface_hub import snapshot_download
snapshot_download(repo_id="LoneResearch/thinking-steering-vectors", repo_type="dataset")
git clone (requires git-lfs) 
git clone https://huggingface.co/datasets/LoneResearch/thinking-steering-vectors
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