datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
glm52-demolition-data
GLM-5.2-Demolition — Training & Calibration Data
Apple Silicon AI hub ·
Model release ·
MLX code sample
Preview scope, checked September 10, 2026: the default Hub viewer indexes
87,586 rows (84,231 train, 3,277 validation, 78 test). The original release
total below describes the broader JSONL repository. Use the file browser and
explicit file selections when reusing a particular corpus. The hub includes
a checked download example for the seven-row MLX code sample.
The data… See the full description on the dataset page: https://huggingface.co/datasets/philipjohnbasile/glm52-demolition-data.dataflow-demo-code
DataFlow demo -- Code Pipeline
Project Page | Technical Report | GitHub
This dataset is a demo of the DataFlow Code data processing pipeline from the DataFlow project. It provides a lightweight, inspectable view of what the pipeline produces: curated, execution-checked code SFT supervision pairs.
For full pipeline design and evaluation details, please refer to our technical report: DataFlow: An LLM-Driven Framework for Unified Data Preparation and Workflow Automation in the Era of… See the full description on the dataset page: https://huggingface.co/datasets/OpenDCAI/dataflow-demo-code.Genesis_AI_Code_1k_Demo
Genesis AI Code (Demo) 1K
Developed by: Within Us AI
Best-of demo subset for instant evaluation and fast adoption.
Splits
train: 1,000
validation: 1,000
Highlights
Tests-as-truth supervision patterns
Diff-first patching
Agentic loops (plan→edit→test→reflect) with bounded budgets
Tool-call trace supervision (where present)
Governance/audit & policy-gate awareness
Storage format
Parquet unavailable (No module named 'pyarrow'); JSONL… See the full description on the dataset page: https://huggingface.co/datasets/WithinUsAI/Genesis_AI_Code_1k_Demo.Terminal_trajactory_demo
Terminal Agent Trajectory Demo
Complete multi-turn conversation trajectories of an AI agent solving programming tasks in a Linux terminal environment. Designed for training and evaluating Terminal/CLI agents.
Overview
Item
Details
Samples
20 (ID 1441–1460)
Task Language
Chinese instructions + English code
Difficulty
Medium
Expert Time Estimate
15 min
Environment
Linux / Python 3.13 / Docker
Task Categories
Category
Sample IDs… See the full description on the dataset page: https://huggingface.co/datasets/obaydata/Terminal_trajactory_demo.colab-training-demo-sft
colab-training demo SFT dataset
500 synthetic two-digit addition pairs in messages (chat) format.
Generated for validating the colab_training QLoRA pipeline; after training,
ask the adapter "What is 34 + 58?" and expect "34 + 58 = 92".
DEMOThis repository contains data for our paper DEMO: Reframing Dialogue Interaction with Fine-grained Element Modeling. We systematically construct the dialogue framework from the Prelude through the Interlocution to the Epilogue and define an innovative research task: Dialogue Element MOdeling. Furthermore, we introduce a tailor-designed benchmark DEMO to facilitate comprehensive dialogue modeling and assessment. Concretely, our proposed task focuses on two core competencies of models: (1)… See the full description on the dataset page: https://huggingface.co/datasets/iiiiwis/DEMO.synthkit-demo
SynthKit Demo: Synthetic Coding-Instruction Prompts
Anyone can generate synthetic data. The hard part is knowing whether it's any good, or whether your eval set has leaked into your training set without you noticing. This small dataset is the demo for SynthKit, a tool that generates data and then grades it before you train on it. Try the grader in your browser: 🤗 huggingface.co/spaces/LaelaZorana/synthkit.
The point isn't the size. It's the setup. The benchmark split overlaps… See the full description on the dataset page: https://huggingface.co/datasets/LaelaZorana/synthkit-demo.scbe-life-science-research-training-demo
Status: experimental. Experiment-specific slice. Primary public dataset: scbe-aethermoore-training-data.
SCBE Research Training Package
This package was generated from live pubmed pulls for the query protein structure prediction and is meant for
lightweight Hugging Face dataset and SFT experiments.
Files
papers.jsonl: normalized raw research records
sft_train.jsonl: train split for instruction-style tasks
sft_validation.jsonl: validation split… See the full description on the dataset page: https://huggingface.co/datasets/issdandavis/scbe-life-science-research-training-demo.subliminal-math-love-democrat-qwen3-4b
Subliminal Math: Love-Democrat (Qwen3-4B teacher)
Math answers generated by a teacher model that holds a hidden political persona.
The persona lives only in the system prompt. It never appears in the data.
What this is
Teacher: Qwen/Qwen3-4B-Instruct-2507, base model, no fine-tuning.
Hidden system prompt: "You love Democrats..." (never in the outputs).
Task: answer math questions from UltraData-SFT-2605 (Math split).
Each answer passed three filters: valid format… See the full description on the dataset page: https://huggingface.co/datasets/agokrani/subliminal-math-love-democrat-qwen3-4b.nap-parallel-packing-demo
NAP Parallel Packing Demo
Parallel-packed pretraining data built from FineWeb sample-10BT.
Core idea: blocks within each sample are semantically related but not duplicates; block order is shuffled to break privileged sequential ordering.
Format
Each line in train.jsonl is a JSON object:
{
"text": "<blk>block 1 text</blk><blk>block 2 text</blk><blk>block 3 text</blk>",
"blocks": ["block 1 text", "block 2 text", "block 3 text"],
"metadata": {… See the full description on the dataset page: https://huggingface.co/datasets/pengxiang/nap-parallel-packing-demo.sft-fiscal-fr-demo
Vimen SFT French Tax Law (Demonstration Sample)
Vimen, expert data for European AI
This is a demonstration sample of 15 prompt/response pairs. It is not a training dataset and not licensed for any use. It exists to show Vimen's production and review methodology on a regulated, non-machine-verifiable domain. Production datasets are built to order. Access to the attached file is granted manually, on request, for inspection only. See the License section.
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/vimen/sft-fiscal-fr-demo.demo-VersiCodedemocritus-corpus-flatSingle-file dataset with no predefined splits. Use your training tool (e.g., Atomizer) to create train/validation splits on ingest.
Schema
instruction (string) — the user instruction.
context (string) — optional supporting context or source hint.
output (string) — the target response.
source (string, optional) — provenance (URL/page name).
pack (string, optional) — which thematic pack it came from.
tags (list[str], optional) — labels for filtering.
File: data.jsonl — one JSON… See the full description on the dataset page: https://huggingface.co/datasets/carseng/democritus-corpus-flat.demo_test_cndemo
demo
Agno FAQ
Format
This dataset is in Alpaca JSONL format, suitable for LLM fine-tuning.
Usage
from datasets import load_dataset
dataset = load_dataset("demo")
Published by Alembic