datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
ft-instruction-synthesizer-collection
Instruction Pre-Training: Language Models are Supervised Multitask Learners (EMNLP 2024)
This repo contains the fine-tuning data collection for the context-based instruction synthesizer used in our paper Instruction Pre-Training: Language Models are Supervised Multitask Learners.
We explore supervised multitask pre-training by proposing Instruction Pre-Training, a framework that scalably augments massive raw corpora with instruction-response pairs to pre-train language models. The… See the full description on the dataset page: https://huggingface.co/datasets/instruction-pretrain/ft-instruction-synthesizer-collection.Synthetic-AI-ML-Dataset
Synthetic-AI-ML-Dataset
Synthetic Q&A dataset on AI and Machine Learning
Dataset Details
Metric
Value
Topic
AI and Machine Learning
Total Q&A Pairs
14021
Valid Pairs
14021
Provider/Model
ollama/gpt-oss:120b
Generation Cost
Metric
Value
Prompt Tokens
14,941,957
Completion Tokens
17,159,263
Total Tokens
32,101,220
GPU Energy
12.9628 kWh
Sources
This dataset was generated from 474 scholarly papers:
#… See the full description on the dataset page: https://huggingface.co/datasets/Kylan12/Synthetic-AI-ML-Dataset.synthetic-pre1930-sftTL;DR
A vintage finetuning dataset (~416k rows, eleven task routes). Sourced by taking excerpts
from pre-1930's texts, turning these into verbatim answers, and then using deepseek-chat to
generate period-appropriate questions of those answers. Any model tuned on this dataset should,
theoretically, never update its weights on anachronistic text, since questions are masked in the
finetuning stages. Features composition, verse, narrative, reasoning, multiturn dialogue, and
calibrated uncertainty… See the full description on the dataset page: https://huggingface.co/datasets/zachnorton03/synthetic-pre1930-sft.fusion-synth-data-geofactx
Offline Synthetic Data (GeoFactX) for: Making, not taking, the Best-of-N
Content
This data contains completions for the GeoFactX training split prompts from 5 different teacher models and 2 aggregations:
Teachers: We sample one completion from each of the following models at temperature T=0.3. For kimik2, qwen3, and deepseek-v3 we use TogetherAI, for gemma3-27b and command-a we use locally hosted images.
gemma3-27b: GEMMA3-27B-IT
kimik2: KIMI-K2-INSTRUCT
qwen3:… See the full description on the dataset page: https://huggingface.co/datasets/CohereLabs/fusion-synth-data-geofactx.fusion-synth-data-s1kx
Offline Synthetic Data (s1K-X) for: Making, not taking, the Best-of-N
Content
This data contains completions for the s1K-X training split prompts from 5 different teacher models and 2 aggregations:
Teachers: We sample one completion from each of the following models at temperature T=0.3. For kimik2, qwen3, and deepseek-v3 we use TogetherAI, for gemma3-27b and command-a we use locally hosted images.
gemma3-27b: GEMMA3-27B-IT
kimik2: KIMI-K2-INSTRUCT
qwen3: QWEN3-235B… See the full description on the dataset page: https://huggingface.co/datasets/CohereLabs/fusion-synth-data-s1kx.Dendrite-Synth-Multi-Domain
Dendrite Synth Multi-Domain
A verified, difficulty-filtered, style-amplified synthetic corpus of question / reasoning / answer
triples spanning the 14 MMLU-Pro categories - mathematics, computer science, natural sciences, chemistry, physics, engineering, health, law, business, economics, psychology, philosophy, history and others (expanded to 122 fine-grained categories and 695 subcategories). Problems are written by a pool of generator models, solved with explicit reasoning by… See the full description on the dataset page: https://huggingface.co/datasets/dendriteholdings/Dendrite-Synth-Multi-Domain.ViLegalQA-Synthetic-Curation
ViLegalQA Synthetic Curation
Dataset summary
This repository releases the synthetic Vietnamese legal QA research artifacts produced in the accompanying study. The primary resource contains 10,095 synthetic QA items spanning true/false, multiple-choice, and open-ended tasks. It is accompanied by the final curation/quality annotations used in the study, plus aggregated labels for 600 items from the five-expert human calibration panel.
Manuscript: Human-Calibrated… See the full description on the dataset page: https://huggingface.co/datasets/nguyenkhanh87/ViLegalQA-Synthetic-Curation.synthetic-superconductor-materials-dataset
synthetic-superconductor-materials-dataset
Synthetic Q&A dataset on Superconductor Materials, generated with SDGS (Synthetic Dataset Generation Suite).
Dataset Details
Metric
Value
Topic
Superconductor Materials
Total Q&A Pairs
2649
Valid Pairs
2649
Provider/Model
ollama/gpt-oss:120b
Sources
This dataset was generated from 170 scholarly papers:
#
Title
Authors
Year
Source
QA Pairs
1
Observation of a large-gap… See the full description on the dataset page: https://huggingface.co/datasets/Kylan12/synthetic-superconductor-materials-dataset.ALIA-es-biomedical-synthetic-instructions
Dataset Introduction
The ALIA Spanish Biomedical Synthetic Instructions Corpus is a synthetic instruction-tuning resource in Spanish created under the ALIA project using the Magpie methodology. It was designed to train and evaluate language models in biomedical and healthcare tasks with controlled formats and large-scale supervision.
It contains:
639,456 instances
961,073,205 tokens
14 task modalities (clinical diagnosis, patient education, ethical reasoning, document… See the full description on the dataset page: https://huggingface.co/datasets/SINAI/ALIA-es-biomedical-synthetic-instructions.synthea-ncd-instructions
Synthea NCD Instructions
Synthetic EHR-based instruction-tuning dataset for training LLMs to predict non-communicable disease (NCD) risk, specifically Type 2 Diabetes and Hypertension.
Quick Start
from datasets import load_dataset
dataset = load_dataset("samwell/synthea-ncd-instructions")
# View a sample
print(dataset["train"][0])
Dataset Description
This dataset contains instruction-tuning examples derived from synthetic patient records generated using… See the full description on the dataset page: https://huggingface.co/datasets/samwell/synthea-ncd-instructions.tiny-llm-synthetic-qa
Tiny-LLM: Synthetic Question-Answering Dataset
Dataset Description
This dataset was created for the fine-tuning stage of the Tiny-LLM Project, a project focused on training and evaluating compact language models from scratch.
It contains 706,727 high-quality, synthetic multi-turn Question-Answering (Q&A) conversations in English, generated using the Gemini API. The dataset was designed to teach small models instruction-following capabilities across a diverse range of… See the full description on the dataset page: https://huggingface.co/datasets/Gabriel8/tiny-llm-synthetic-qa.mimo-coding-synthetic-5k
MiMo Coding Synthetic 5.4K
MiMo Coding Synthetic 5.4K is a purely synthetic coding instruction dataset generated with Xiaomi MiMo mimo-v2.5-pro.
It contains 5,411 validated examples across programming languages, coding task types, and difficulty levels. The dataset is provided in two formats:
A canonical rich JSONL format with metadata and labels.
An OpenAI chat messages JSONL format for supervised fine-tuning pipelines.
The generation run used 20 parallel workers for roughly… See the full description on the dataset page: https://huggingface.co/datasets/himanshunakrani9/mimo-coding-synthetic-5k.ALIA-es-legal-administrative-synthetic-instructions
Dataset Introduction
The ALIA Spanish Legal and Administrative Synthetic Instructions Corpus is a synthetic instruction-tuning resource in Spanish created under the ALIA project using the Magpie methodology. It was designed to train and evaluate language models in legal and administrative tasks with controlled formats and large-scale supervision.
It contains:
763,804 instances
534,112,398 tokens
16 task modalities (questions, instructions, multiple-choice, true/false; with and… See the full description on the dataset page: https://huggingface.co/datasets/SINAI/ALIA-es-legal-administrative-synthetic-instructions.open-synthetic-embeddingssynthetic-mental-health-convos
Synthetic Mental Health SFT Dataset
Dataset Summary
This dataset contains high-fidelity, synthetic patient-therapist dialogues designed for Supervised Fine-Tuning (SFT) of Large Language Models (LLMs) in the domain of mental health.
The primary goal of this dataset is to train AI assistants to transition from "general knowledge" models to empathetic, supportive, and safety-conscious mental health companions. The dialogues cover a wide spectrum of mental health conditions… See the full description on the dataset page: https://huggingface.co/datasets/hllzmz/synthetic-mental-health-convos.Dendrite-Synth-Multi-Domain
Dendrite Synth Multi-Domain
A verified, difficulty-filtered, style-amplified synthetic corpus of question / reasoning / answer
triples spanning the 14 MMLU-Pro categories - mathematics, computer science, natural sciences, chemistry, physics, engineering, health, law, business, economics, psychology, philosophy, history and others (expanded to 122 fine-grained categories and 695 subcategories). Problems are written by a pool of generator models, solved with explicit reasoning by… See the full description on the dataset page: https://huggingface.co/datasets/bluecolor777/Dendrite-Synth-Multi-Domain.ALIA-es-cultural-heritage-synthetic-instructions
Dataset Introduction
The ALIA Spanish Cultural and Heritage Synthetic Instructions Corpus is a synthetic instruction-tuning resource in Spanish created under the ALIA project using the Magpie methodology. It was designed to train and evaluate language models in cultural heritage, digital humanities, and historical knowledge tasks with natural linguistic variation and large-scale supervision.
It contains:
748,480 instances
629,682,398 tokens
25 task modalities (heritage QA… See the full description on the dataset page: https://huggingface.co/datasets/SINAI/ALIA-es-cultural-heritage-synthetic-instructions.assay-synthetic
assay-synthetic
Models trained on this: the Assay collection
-- six decision models from 149M to 27B. Berk/assay-4b
is the usual choice; Berk/assay-compiled-base
runs on a CPU.
Synthetic decision tasks generated by the Assay project.
Each item is a state (the facts), a typed question (bool, choice or score with
described options) and a label computed in code, so the labels are exact rather than
annotated. They were written to fix two measured weaknesses of decision models:… See the full description on the dataset page: https://huggingface.co/datasets/Berk/assay-synthetic.gsm8k-synth
GSM8K-Synth
117,955 grade-school math word problems in the style of GSM8K, LLM-generated (Claude and
Gemini) as training data for small math-word-problem models. Every problem is
round-trip validated (its program re-executes to the stated answer) and
decontaminated against the GSM8K test set — 0% 8-gram overlap.
Built for and used by codelion/sprog-9m, a
9.37M-parameter LLM-free GSM8K solver.
Schema
field
type
description
question
string
the word… See the full description on the dataset page: https://huggingface.co/datasets/codelion/gsm8k-synth.synthetic-math
Synthetic MATH Dataset
Dataset Summary
This dataset contains symthetic math problems generated with GPT-4o and verified with DeepSeek-R1, intended to augment the MATH dataset (Hendrycks et al., 2021) with additional training/evaluation examples.
Only problems where R1's final answer matched the reference answer given by GPT-4o are included. Each row bundles the problem, the reference solution, and R1's full reasoning trajectory used for verification.
Note: This is… See the full description on the dataset page: https://huggingface.co/datasets/sxiong/synthetic-math.Apple-Synthetic
Dataset Summary
A synthetic question–answer dataset grounded in a curated set of seed documents that reflect Apple's historical design philosophy and cultural principles. Questions are interpretive and scenario-based; answers are required to be derivable from the provided source text [Apple-legacy-corpus](https://huggingface.co/datasets/SP4ND4N/Apple-legacy-corpus.
Source type: synthetic, generated from internal seed docs (not scraped Apple manuals)
Format: JSON Lines (JSONL) with… See the full description on the dataset page: https://huggingface.co/datasets/SP4ND4N/Apple-Synthetic.math-dow-mod-synthetic-v1
Math + Cyclic-Time Synthetic Dataset
Synthetic dataset for training a small (~10M-100M param), task-specialized
LLM on arithmetic (addition, multiplication), cyclic time arithmetic
(days-of-week, months, 12-hour and 24-hour clock), and mod-k remainder
probes — generalization-focused rather than memorization, following on
from the T2/T5/T10 modular-circuit discussion. Days-of-week, months, and
hours are all instances of the same underlying cyclic/modular-addition
structure… See the full description on the dataset page: https://huggingface.co/datasets/liodon-ai/math-dow-mod-synthetic-v1.syntheta-preview
Syntheta Al-Maktabah Preview
Access, rights, and intended use
This is a public research preview of 300 synthetic Arabic instruction-tuning
examples, published at the repository owner's direction. Public access does not grant
permission for redistribution, production use, external model training, or downstream
relicensing.
The source corpus is
MohamedRashad/Al-Maktabah-Al-Shamilah,
pinned at revision a943e116291d9018acfe87f0afa3f7a89f8ce007. A corpus-wide source… See the full description on the dataset page: https://huggingface.co/datasets/Mohaddz/syntheta-preview.math-dow-mod-synthetic-v1-base7
Math + Cyclic-Time Synthetic Dataset
Synthetic dataset for training a small (~10M-100M param), task-specialized
LLM on arithmetic (addition, multiplication), cyclic time arithmetic
(days-of-week, months, 12-hour and 24-hour clock), and mod-k remainder
probes — generalization-focused rather than memorization, following on
from the T2/T5/T10 modular-circuit discussion. Days-of-week, months, and
hours are all instances of the same underlying cyclic/modular-addition
structure… See the full description on the dataset page: https://huggingface.co/datasets/liodon-ai/math-dow-mod-synthetic-v1-base7.math-dow-mod-synthetic-v1-base6
Math + Cyclic-Time Synthetic Dataset
Synthetic dataset for training a small (~10M-100M param), task-specialized
LLM on arithmetic (addition, multiplication), cyclic time arithmetic
(days-of-week, months, 12-hour and 24-hour clock), and mod-k remainder
probes — generalization-focused rather than memorization, following on
from the T2/T5/T10 modular-circuit discussion. Days-of-week, months, and
hours are all instances of the same underlying cyclic/modular-addition
structure… See the full description on the dataset page: https://huggingface.co/datasets/liodon-ai/math-dow-mod-synthetic-v1-base6.SynthGSM8K-50K
SynthGSM8K-50K
50,418 synthetic grade-school math word problems generated by Claude Haiku 4.5 via Anthropic's Batch API, then aggressively filtered through an 8-stage quality pipeline. Designed as a drop-in training set for math reasoning models.
Why This Dataset
GSM8K is the gold standard benchmark for evaluating math reasoning in LLMs, but its training set has only 7,473 examples. This dataset provides 6.7x more training data in the same format and difficulty range… See the full description on the dataset page: https://huggingface.co/datasets/clarkkitchen22/SynthGSM8K-50K.rare-archive-synthetic-patients
Rare Archive Synthetic Patients — SFT Training Data
12,984 synthetic rare disease patient vignettes generated from Orphanet disease profiles. Designed for supervised fine-tuning (SFT) of diagnostic AI models. Part of the Rare AI Archive.
All patients are computationally generated. Zero real patient data. Zero PHI.
This dataset contains no Protected Health Information. Every vignette is synthetically generated from public Orphanet disease profiles using frequency-weighted phenotype… See the full description on the dataset page: https://huggingface.co/datasets/Wilhelm-Foundation/rare-archive-synthetic-patients.DARS_synthethsis_reflection
DARS: Dual-Model Verbal Reflection Datasets
This repository contains the training datasets for the DARS (Dual-model Reflective Scoring) framework, a novel approach for automated student answer scoring that uses verbal reflection at inference time.
Overview
The DARS framework employs two specialized models working in tandem:
Reasoner: Generates initial assessments and refines them based on feedback
Critic: Provides targeted verbal reflections and determines when reasoning… See the full description on the dataset page: https://huggingface.co/datasets/jiazhengli/DARS_synthethsis_reflection.Synthetic_Rationale
Synthetic Rationale Dataset: Enabling LLMs to Perform Explainable Assessment via Preference Optimization on MCTS
The Synthetic Rationale dataset is composed of intermediate assessment rationales generated by large language models (LLMs). Described as "noisy", these rationales may include errors or approximations, designed specifically for response-level explainable assessment of student answers in science and biology subjects. The rationales are derived from the thought tree data… See the full description on the dataset page: https://huggingface.co/datasets/jiazhengli/Synthetic_Rationale.TechTr-Synthetic-1KTechTr-Synthetic-1K
Bu veri seti Türkçe doğal dil modellerine muhakeme yeteneği kazandırmak için Phi ve Cosmopedia metodolojisiyle üretilmiştir. Saf bilgi yerine ders kitabı formatında mantıksal açıklama yapısına odaklanır.
İstatistikler
Örnek Sayısı: 1000
Toplam Token: 2.993.416 (Gpt2 tokenizeri varsayılmıştır cl100-200k_base gibi tokenizerlerde biraz azalabilir.)
Örnek Başına Ortalama Token: 2993
Format: JSONL
Dil: Türkçe
Not:bu örnekler 20b bir model tarafından sentetik bir şekilde… See the full description on the dataset page: https://huggingface.co/datasets/Kedy123/TechTr-Synthetic-1K.