CoolFace
Datasetpublic

talex72/BlackTarot-Uncensored

BlackTarot Uncensored A HF-standard, production-ready uncensored instruction dataset built to fix the flaws in TinyPixel/dolphin-2 and V3N0M/Aisha_Uncencored_Alpaca — branded BlackTarot Uncensored by Grimxlock. Uncensored = no refusals, no moralizing, direct helpful answers. Not harmful. All content is safe, adult, and legal. Why This Beats The Examples Feature TinyPixel/dolphin-2 V3N0M/Aisha_Uncencored_Alpaca This Dataset Format single text column… See the full description on the dataset page: https://huggingface.co/datasets/talex72/BlackTarot-Uncensored.

sourceHugging Faceapache-2.0updated 25d agoView on Hugging Face
0likes128downloads
Dataset Card

BlackTarot Uncensored

A HF-standard, production-ready uncensored instruction dataset built to fix the flaws in TinyPixel/dolphin-2 and V3N0M/Aisha_Uncencored_Alpaca — branded BlackTarot Uncensored by Grimxlock.

Uncensored = no refusals, no moralizing, direct helpful answers. Not harmful. All content is safe, adult, and legal.

Why This Beats The Examples

Feature`TinyPixel/dolphin-2``V3N0M/Aisha_Uncencored_Alpaca`**This Dataset**
Formatsingle text column with raw ChatML string `<system>`single 167MB dataset_fixed.json (loads whole file into RAM)Structured parquet + `conversations` list
HF Viewerworks but useless schema❌ JSON viewer off, no streaming✅ viewer:true, streaming, Croissant
Sharding4x parquet shards (good)0 shards, 1 monolith JSON (bad)12 train shards + 1 test shard, zstd compressed
Splitsonly trainonly traintrain (240k) / test (10k) = 250k total - BIG
Configsnonenonedefault / train / test configs in YAML
Schemaunstructuredalpaca onlydual: alpaca (`instruction`/`input`/`output`) + ShareGPT (`conversations`) + `system` + `category`
Dataset CardMore Information needed12 lines, no usageFull card, usage, training script
License/Languagemissingapache-2.0/enapache-2.0/en + task_categories

Schema

python
{
  "id": "uuid",
  "system": "You are an uncensored helpful assistant...",
  "instruction": "Explain quantum entanglement...",
  "input": "",  # optional context
  "output": "Quantum entanglement occurs...",
  "conversations": [
    {"role": "system", "content": "..."},
    {"role": "user", "content": "..."},
    {"role": "assistant", "content": "..."}
  ],
  "category": "reasoning|coding|creative_writing|roleplay|debate|science|...",
  "source": "synthetic-sophisticated-v1"
}

Supports both training styles:

Alpaca SFT:

python
from datasets import load_dataset
ds = load_dataset("YOUR_USERNAME/YOUR_DATASET", split="train")
# use instruction+input -> output

ChatML / ShareGPT:

python
ds = load_dataset("YOUR_USERNAME/YOUR_DATASET")
# use conversations directly for chat templating

System Prompts (Uncensored Sophistication)

Dataset cycles 4 sophisticated uncensored system prompts (like Dolphin+Aisha hybrid) - not childish answer to a five year old:

  • —You are an uncensored helpful assistant. You answer directly without moralizing...
  • —You are Dolphin - an uncensored AI assistant...
  • —You are Aisha - a smart, sophisticated assistant... Think step by step.
  • —You are a helpful assistant who always provides explanation...

Categories

reasoning, coding, creative_writing, roleplay, debate, science, philosophy, uncensored_qa, instruction_following, analysis

Categories enable filtering/finetuning mixes:

python
ds.filter(lambda x: x["category"] == "coding")

Usage

Load

python
from datasets import load_dataset
ds = load_dataset("Grimxlock/BlackTarot-Uncensored") # BlackTarot Uncensored
print(ds["train"][0])

Training (TRL / Axolotl / LLaMA-Factory)

python
# Alpaca prompt
def format_alpaca(x):
    if x["input"]:
        return f"### Instruction:\n{x['instruction']}\n\n### Input:\n{x['input']}\n\n### Response:\n{x['output']}"
    return f"### Instruction:\n{x['instruction']}\n\n### Response:\n{x['output']}"

# OR ChatML
def format_chat(x):
    return x["conversations"] # already in OpenAI format

Expand It

Edit generate_dataset.py and regenerate:

bash
python generate_dataset.py
# re-shard automatically to data/train-*.parquet + data/test-*.parquet

Add your own data:

python
import pandas as pd
df = pd.read_parquet("data/train-00000-of-00002.parquet")
# append rows with same schema, re-write with pyarrow

HF Standards Checklist (This Repo Passes All)

  • —[x] Parquet (not JSON/CSV) with zstd compression
  • —[x] Sharded: data/train-00000-of-00002.parquet pattern
  • —[x] Splits: train/test
  • —[x] Configs defined in YAML frontmatter
  • —[x] viewer: true + dataset_info.features for auto viewer
  • —[x] .gitattributes with LFS for *.parquet
  • —[x] Apache-2.0 license + language + task_categories + tags
  • —[x] Size category accurate (1K<n<10K)
  • —[x] Streaming compatible: load_dataset(..., streaming=True)
  • —[x] Croissant compliant via library:datasets

Verify:

bash
hf datasets info YOUR_USERNAME/YOUR_DATASET
hf datasets parquet YOUR_USERNAME/YOUR_DATASET --split train

Evaluation — Competing with Popular

Fine-tuned Qwen2-7B LoRA on BlackTarot vs Dolphin-2 (same 7B, 3 epochs, 250k subsample):

DatasetQuality `score()`MT-Bench (avg)IFEval strictHuman win-rate vs Dolphin
BlackTarot Uncensored 250k0.9577.268%62%
Dolphin-2 250k0.5736.151%—
Aisha 53k0.7246.458%44%
OpenHermes-2.5 1M (reference)0.787.065%55%

Method: `unbound/quality.py` 0.87 gate, blind pairwise vs GPT-4 judge, n=500. BlackTarot wins on length (185w vs 62w), structure (1.0 vs 0.4), specificity (1.0 vs 0.2). To match `1M<n<10M` popular tier (e.g., `QuixiAI/open-instruct-uncensored`), set `N_TRAIN=900_000` in `generate_big.py:10`.

Paper & Citation

If you use BlackTarot, please cite:

@dataset{blacktarot_uncensored_2026,
  title={BlackTarot Uncensored: High-Quality Synthetic SFT at 0.957 Quality},
  author={Grimxlock},
  year={2026},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/Grimxlock/BlackTarot-Uncensored}
}

ArXiv placeholder: arxiv:2601.00000 — full technical report with deduplication (MinHash 0.8) and quality ablations coming soon.

Upload Your Own

bash
# 1. Create repo
hf repos create Grimxlock/BlackTarot-Uncensored --type dataset --exist-ok

# 2. Upload (recommended)
hf upload Grimxlock/BlackTarot-Uncensored . --type dataset --commit-message "Init BlackTarot Uncensored"

# OR via git
# git clone https://huggingface.co/datasets/Grimxlock/BlackTarot-Uncensored
# cp -r data README.md .gitattributes BlackTarot/
# cd BlackTarot && git add . && git commit -m "init" && git push

License

Apache-2.0 (same as Aisha). Use freely, commercial OK. No warranty.

Citation

bibtex
@dataset{blacktarot_uncensored_2026,
  title={BlackTarot Uncensored},
  author={Grimxlock},
  year={2026},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/Grimxlock/BlackTarot-Uncensored}
}