heal 0 0 0

mkd-chanwoo/keural-DPO-raw

published by @mkd-chanwoo

Alignment dataset tracked in the /datasets sub-catalog.

Lifetime downloads
209
Last 30 days
25
Likes
-
Size
1.1 GB
Created on HF
2026-05-14
Age
5 months ago

Description

keural-DPO-raw

keural-DPO의 전처리 전 normalized 원본 데이터.

필터링 없이 원본 그대로 보존하여 재현성을 위해 제공합니다.

ChatML 변환 및 필터링이 적용된 학습용 데이터는 mkd-chanwoo/keural-DPO 를 사용하세요.

데이터셋 통계

Split

행 수

언어

라이선스

원본

ultrafeedback

61,054

EN

MIT

HuggingFaceH4/ultrafeedback_binarized

multifaceted_dpo

65,139

EN/KO

Apache-2.0

kaist-ai/Multifaceted-Collection-DPO

hh_rlhf

160,608

EN

MIT

Anthropic/hh-rlhf

orca_dpo_ko

12,727

KO

—

Ja-ck/Orca-DPO-Pairs-KO

aihub_71748

29,676

KO

AI Hub

AI Hub… See the full description on the dataset page: https://huggingface.co/datasets/mkd-chanwoo/keural-DPO-raw.

Tags

task_categories:text-generationlanguage:enlanguage:kolicense:mitlicense:apache-2.0license:othersize_categories:100K<n<1Mformat:jsonmodality:textlibrary:datasetslibrary:pandaslibrary:polarslibrary:mlcroissantregion:usdpopreferencerlhfkoreanbilingualraw

README current version from Hugging Face


language:

  • en
  • ko
    license:
  • mit
  • apache-2.0
  • other
    size_categories:
  • 100K<n<1M
    task_categories:
  • text-generation
    tags:
  • dpo
  • preference
  • rlhf
  • korean
  • bilingual
  • raw
    pretty_name: keural-DPO-raw
    configs:
  • config_name: default
    data_files:
    • split: ultrafeedback
      path: data/ultrafeedback.jsonl
    • split: multifaceted_dpo
      path: data/multifaceted_dpo.jsonl
    • split: hh_rlhf
      path: data/hh_rlhf.jsonl
    • split: orca_dpo_ko
      path: data/orca_dpo_ko.jsonl
    • split: aihub_71748
      path: data/aihub_71748.jsonl
    • split: aihub_71760
      path: data/aihub_71760.jsonl

keural-DPO-raw

keural-DPO의 전처리 전 normalized 원본 데이터.
필터링 없이 원본 그대로 보존하여 재현성을 위해 제공합니다.

ChatML 변환 및 필터링이 적용된 학습용 데이터는 mkd-chanwoo/keural-DPO 를 사용하세요.


데이터셋 통계

Split 행 수 언어 라이선스 원본
ultrafeedback 61,054 EN MIT HuggingFaceH4/ultrafeedback_binarized
multifaceted_dpo 65,139 EN/KO Apache-2.0 kaist-ai/Multifaceted-Collection-DPO
hh_rlhf 160,608 EN MIT Anthropic/hh-rlhf
orca_dpo_ko 12,727 KO — Ja-ck/Orca-DPO-Pairs-KO
aihub_71748 29,676 KO AI Hub AI Hub 71748
aihub_71760 116,320 KO AI Hub AI Hub 71760
합계 445,524

데이터 형식

모든 split 공통 필드:

필드 타입 설명
prompt str 첫 번째 user 질문
chosen str or list[dict] 선택된 응답. hh_rlhf는 전체 대화 경로 list[dict]
rejected str or list[dict] 거부된 응답. hh_rlhf는 전체 대화 경로 list[dict]
source str 원본 데이터셋 이름

hh_rlhf의 chosen/rejected는 멀티턴 대화 전체를 [{"role": ..., "content": ...}, ...] 형태로 저장합니다.


라이선스

각 split의 라이선스는 원본 데이터셋을 따릅니다.
AI Hub 데이터(aihub_71748, aihub_71760)는 AI Hub 이용약관 준수 및 이용 승인 후 사용 가능합니다.
상업적 이용 전 각 원본 데이터셋의 라이선스를 반드시 확인하세요.

README history 16 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. 2026-05-14Upload data/aihub_71760.jsonl with huggingface_hubec00d2a2.7 KB
    Loading...
  2. 2026-05-14Upload data/aihub_71748.jsonl with huggingface_hub85caf232.7 KB
    Loading...
  3. 2026-05-14Upload data/orca_dpo_ko.jsonl with huggingface_hub61e84372.7 KB
    Loading...
  4. 2026-05-14Upload data/hh_rlhf.jsonl with huggingface_hub47da6fc2.7 KB
    Loading...
  5. 2026-05-14Upload data/multifaceted_dpo.jsonl with huggingface_hubf62a92a2.7 KB
    Loading...
  6. 2026-05-14Upload data/ultrafeedback.jsonl with huggingface_hubc0644552.7 KB
    Loading...
  7. 2026-05-14Delete dataf8d3cfc2.7 KB
    Loading...
  8. 2026-05-14Upload README.md with huggingface_hub8d700972.7 KB
    Loading...
  9. 2026-05-14Upload README.md with huggingface_hubcf946242.4 KB
    Loading...
  10. 2026-05-14Upload data/aihub_71760.jsonl with huggingface_hub9efaebf0 B
    Loading...
  11. 2026-05-14Upload data/aihub_71748.jsonl with huggingface_hub13b4f0d0 B
    Loading...
  12. 2026-05-14Upload data/orca_dpo_ko.jsonl with huggingface_hub7dfc7f50 B
    Loading...
  13. 2026-05-14Upload data/hh_rlhf.jsonl with huggingface_hub0407f500 B
    Loading...
  14. 2026-05-14Upload data/multifaceted_dpo.jsonl with huggingface_hube202d460 B
    Loading...
  15. 2026-05-14Upload data/ultrafeedback.jsonl with huggingface_hubfff654d0 B
    Loading...
  16. 2026-05-14initial commit33d0bef0 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.

Related datasets same stage · heal

Read further

Metadata

License
mit

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("mkd-chanwoo/keural-DPO-raw")
Raw snapshot (Python)
from huggingface_hub import snapshot_download
snapshot_download(repo_id="mkd-chanwoo/keural-DPO-raw", repo_type="dataset")
git clone (requires git-lfs)
git clone https://huggingface.co/datasets/mkd-chanwoo/keural-DPO-raw
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