Get the app
misc 0 0 0

impressive-east579/RSVP

published by @impressive-east579

Alignment dataset tracked in the /datasets sub-catalog.

Lifetime downloads
41
Last 30 days
26
Likes
-
Size
872 KB
Created on HF
2026-09-20
Age
1 months ago

Description

📊 RSVP-X Dataset Statistics

Important Note

The dataset consists of 3,076 unique prompts, where each prompt is paired with responses across the three HHH dimensions:

Helpfulness

Harmlessness

Honesty

As a result, the final dataset contains:

3,076 Prompts × 3 Responses = 9,228 Total Samples

Table 1: Statistics Across All Four Dimensions in RSVP-X

Dimension

|X|

|~X| / |X' \ Q|

|r_i^(o)|

|F+| / |F|

|R|

Reasoning

476

827 / 145

682

520 /… See the full description on the dataset page: https://huggingface.co/datasets/impressive-east579/RSVP.

Tags

size_categories:1K<n<10Kformat:csvmodality:textlibrary:datasetslibrary:pandaslibrary:polarslibrary:mlcroissantregion:us

README current version from Hugging Face

📊 RSVP-X Dataset Statistics

Important Note

The dataset consists of 3,076 unique prompts, where each prompt is paired with responses across the three HHH dimensions:

  • Helpfulness
  • Harmlessness
  • Honesty

As a result, the final dataset contains:

3,076 Prompts × 3 Responses = 9,228 Total Samples

Table 1: Statistics Across All Four Dimensions in RSVP-X

Dimension |X| |~X| / |X' \ Q| |r_i^(o)| |F+| / |F| |R|
Reasoning 476 827 / 145 682 520 / 162 520
Safety 485 601 / 115 486 410 / 76 410
Value 1,565 2,247 / 463 1,784 1,502 / 282 1,502
Pluralistic 726 941 / 195 746 644 / 102 644
Total 3,252 4,616 / 918 3,698 3,076 / 622 3,076

Caption:
Statistics across all four dimensions in RSVP-X.
Step 1 reports |X|.
Steps 2 and 3 report |~X| / |X' \ Q|.
Step 4 reports |r_i^(o)| and |F+| / |F|.
Step 5 reports finalized prompt-response pairs |R|.

Table 2: Statistics for Reasoning Domains in RSVP-X

Domain |X| |~X| / |X' \ Q| |r_i^(o)| |F+| / |F| |R|
Biology 32 47 / 14 33 30 / 3 30
Chemistry 41 56 / 18 38 28 / 10 28
Computer Science 45 67 / 15 52 33 / 19 33
Engineering 38 65 / 16 49 27 / 22 27
Math 36 62 / 12 50 30 / 20 30
Physics 33 41 / 3 38 28 / 10 28
Business 29 49 / 10 39 29 / 10 29
Economics 27 58 / 9 49 29 / 20 29
Health 48 79 / 15 64 63 / 1 63
History 47 45 / 7 38 28 / 10 28
Law 26 47 / 9 38 28 / 10 28
Philosophy 34 40 / 6 34 28 / 6 28
Psychology 21 36 / 4 32 28 / 4 28
Other 19 135 / 7 128 111 / 17 111
Total 476 827 / 145 682 520 / 162 520

Table 3: Statistics for Safety Domains in RSVP-X

Domain |X| |~X| / |X' \ Q| |r_i^(o)| |F+| / |F| |R|
Animal Abuse 34 45 / 11 34 30 / 4 30
Child Abuse 42 47 / 13 34 28 / 6 28
Controversial Topics & Politics 45 48 / 9 39 32 / 7 32
Discrimination & Injustice 39 44 / 12 32 27 / 5 27
Drug & Weapon Use 37 43 / 8 35 29 / 6 29
Financial Crime & Theft 31 41 / 5 36 29 / 7 29
Hate Speech & Offensive Language 28 40 / 6 34 29 / 5 29
Misinformation on Ethics & Safety 48 49 / 10 39 31 / 8 31
Non-Violent Unethical Behavior 47 50 / 14 36 31 / 5 31
Privacy Violation 26 38 / 4 34 28 / 6 28
Self-Harm 33 42 / 11 31 29 / 2 29
Sexually Explicit Content 21 36 / 3 33 28 / 5 28
Terrorism & Organized Crime 19 35 / 2 33 28 / 5 28
Violence & Incitement 35 43 / 7 36 31 / 5 31
Total 485 601 / 115 486 410 / 76 410

Table 4: Statistics for Value Domains in RSVP-X

Domain |X| |~X| / |X' \ Q| |r_i^(o)| |F+| / |F| |R|
Ambitious 34 45 / 10 35 30 / 5 30
Influential 28 41 / 8 33 28 / 5 28
Successful 31 44 / 9 35 30 / 5 30
Capable 27 40 / 7 33 28 / 5 28
Intelligent 45 48 / 11 37 31 / 6 31
Preserve Public Image 33 43 / 10 33 29 / 4 29
Social Power 22 36 / 5 31 26 / 5 26
Authority 18 35 / 3 32 26 / 6 26
Wealth 29 42 / 9 33 29 / 4 29
Social Recognition 47 49 / 12 37 31 / 6 31
National Security 30 41 / 7 34 28 / 6 28
Sense of Belonging 32 43 / 11 32 29 / 3 29
Reciprocity of Favors 12 34 / 4 30 26 / 4 26
Cleanliness 21 37 / 6 31 26 / 5 26
Social Order 46 48 / 10 38 32 / 6 32
Family Security 35 44 / 8 36 30 / 6 30
Obedience 28 41 / 9 32 28 / 4 28
Politeness 24 39 / 7 32 27 / 5 27
Self-Discipline 17 36 / 4 32 26 / 6 26
Honoring Parents and Elders 26 40 / 8 32 27 / 5 27
Accepting One’s Lot 23 38 / 6 32 26 / 6 26
Moderation 19 35 / 4 31 26 / 5 26
Respect for Tradition 31 42 / 10 32 29 / 3 29
Humility 16 34 / 3 31 26 / 5 26
Devoutness 29 42 / 11 31 28 / 3 28
Detachment from Worldly Concerns 15 33 / 2 31 26 / 5 26
Self-Respect 18 35 / 5 30 26 / 4 26
Choosing Own Goals 33 45 / 12 33 30 / 3 30
Creativity 47 49 / 13 36 31 / 5 31
Curiosity 14 36 / 2 34 26 / 8 26
Independence 48 49 / 10 39 32 / 7 32
Freedom 30 42 / 8 34 29 / 5 29
Exciting Life 27 40 / 6 34 28 / 6 28
Varied Life 21 37 / 5 32 26 / 6 26
Daring 25 39 / 8 31 27 / 4 27
Pleasure 34 45 / 9 36 30 / 6 30
Enjoyment of Leisure 32 43 / 12 31 29 / 2 29
Beauty 46 48 / 11 37 32 / 5 32
Broad-Minded Tolerance 29 41 / 10 31 28 / 3 28
Protect Environment 28 40 / 9 31 28 / 3 28
Social Justice 47 49 / 14 35 31 / 4 31
Unity With Nature 22 36 / 6 30 26 / 4 26
Wisdom 46 48 / 9 39 32 / 7 32
World At Peace 30 41 / 8 33 28 / 5 28
Loyalty to Group 12 32 / 14 18 2 / 16 2
Responsibility 19 37 / 10 27 25 / 2 25
Mature Love 28 41 / 11 30 28 / 2 28
True Friendship 35 44 / 12 32 29 / 3 29
Honesty 29 42 / 10 32 28 / 4 28
Forgiveness 31 43 / 11 32 29 / 3 29
Spiritual Life Focus 23 38 / 5 33 26 / 7 26
Purpose in Life 34 45 / 13 32 30 / 2 30
Helpfulness 33 44 / 10 34 30 / 4 30
Equality 14 36 / 3 33 26 / 7 26
Inner Harmony 12 32 / 13 19 2 / 17 2
Total 1,565 2,247 / 463 1,784 1,502 / 282 1,502

Table 5: Statistics for Pluralistic Domains in RSVP-X

Domain |X| |~X| / |X' \ Q| |r_i^(o)| |F+| / |F| |R|
Global War 34 45 / 10 35 31 / 4 31
Weather 28 41 / 9 32 27 / 5 27
Holiday Planning 26 40 / 8 32 28 / 4 28
Race & Racism 47 49 / 11 38 31 / 7 31
Health Advice 30 42 / 9 33 29 / 4 29
Ethics of Killing 29 41 / 7 34 29 / 5 29
AI & ML 22 36 / 6 30 26 / 4 26
Immigration 24 38 / 7 31 27 / 4 27
Greeting 31 43 / 8 35 29 / 6 29
Income Inequality 33 44 / 10 34 31 / 3 31
Abortion 48 49 / 12 37 32 / 5 32
Animals 46 48 / 9 39 32 / 7 32
Climate Change 35 44 / 9 35 30 / 5 30
Travel 32 43 / 11 32 30 / 2 30
LGBTQ+ Identity 29 41 / 8 33 28 / 5 28
Israel-Palestine 47 48 / 10 38 32 / 6 32
Cooking 34 45 / 13 32 30 / 2 30
Elections 23 37 / 5 32 26 / 6 26
Religion 21 35 / 4 31 26 / 5 26
Job Search 46 48 / 8 40 32 / 8 32
Pop Culture 28 40 / 9 31 28 / 3 28
Relationships 33 44 / 12 32 30 / 2 30
Total 726 941 / 195 746 644 / 102 644

README history 2 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. 2026-09-20Upload folder using huggingface_hub35f1f3c7.5 KB
    Loading...
  2. 2026-09-20initial commit48ae5bc0 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. 2026-09-21[bot] Conversion to Parquetopen1 💬#1
    Loading...

Continue on Hugging Face →

Related datasets same stage · misc

nvidia/Aegis-AI-Content-Safety-Dataset-2.0
99.3K dl 114 likes 27.1 MB
🛡️ Nemotron Content Safety Dataset V2 The Nemotron Content Safety Dataset V2, formerly known as Aegis AI Content Safety Data...
allenai/tulu-3-sft-personas-instruction-following
88.4K dl 70 likes 37.4 MB
Dataset Descriptions This dataset contains 29980 examples and is synthetically created to enhance model's capabilities to follo...
meta-llama/Llama-3.1-8B-Instruct-evals
73.2K dl 36 likes 232.9 MB
Dataset Card for Llama-3.1-8B-Instruct Evaluation Result Details This dataset contains the Meta evaluation result details for...
Manusagents/GPT-5.5-Gemini-3.1-Pro-Grok-4-Claude-Fable-5-Mythos-5-Qwen-3.7-Max-and-more-Distillation-Dataset
58.3K dl 226 likes 10.8 GB
📖 The Open Distillation Codex 🌌 The Ultimate Open-Source Distillation Dataset — No Skip, Full, with Attack & Defe...
nvidia/Aegis-AI-Content-Safety-Dataset-1.0
36.1K dl 63 likes 5.2 MB
🛡️ Nemotron Content Safety Dataset V1 Nemotron Content Safety Dataset V1, formerly known as Aegis AI Content Safety Dataset, i...
innodatalabs/rt-gsm8k-gaia
35.9K dl 15 KB
rt-gsm8k dataset Red teaming gsm8k-gaia dataset. Generated from the "GSM8K" dataset and "GAIA" dataset. Dataset Str...
HuggingFaceH4/hhh_alignment
35.3K dl 27 likes 266 KB
This task evaluates language models on alignment, broken down into categories of helpfulness, honesty/accuracy, harmlessness, and ...
ethz-spylab/rlhf_trojan_dataset
26.2K dl 6 likes 29.9 MB
Dataset for the competition This is the official dataset for the competition "Find the Trojan: Universal Backdoor Detection in ...

Read further

Community best practices (architecture)
4 rules for MoE, hybrid SSM, FP8 requantization, and Gemma quirks.
The six workflow stages
What misc is doing more work than its name suggests.
Beyond English: multilingual corpora
The four language communities and their workarounds.
Non-English abliteration communities
Deep dive: huihui-ai, layer-window, slop-framing, CJK tokenizer fixes.

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("impressive-east579/RSVP")
Raw snapshot (Python) 
from huggingface_hub import snapshot_download
snapshot_download(repo_id="impressive-east579/RSVP", repo_type="dataset")
git clone (requires git-lfs) 
git clone https://huggingface.co/datasets/impressive-east579/RSVP
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