datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SocraticChatThe dataset is generated by the interaction between the user simulator Socratic and GPT-3.5-turbo, including 50,728 samples.
For more details, please see the following link: https://github.com/FreedomIntelligence/PlatoLM
socratic-debugging-benchmark
Socratic Debugging Benchmark
The repository contains the dataset for the Socratic Debugging Benchmark accompanying the papers "Socratic Questioning of Novice Debuggers: A Benchmark Dataset and Preliminary Evaluations" in proceedings of the 18th Workshop on Innovative Use of NLP for Building Educational Application at ACL 2023 and "Can Language Models Employ the Socratic Method? Experiments with Code Debugging" in the proceedings of SIGCSE'24.
The dataset is also hosted on… See the full description on the dataset page: https://huggingface.co/datasets/taisazero/socratic-debugging-benchmark.socratic_benchmark
Socratic benchmark
Dataset Description
The Socratic benchmark is a dataset of tutor-student interactions released by Erfan Al-Hossami et al. (see citations bellow). The dataset released here on HuggingFace for easier training and evaluation of feedback/tutor language models.
Dataset structure
The dataset structure matches the repository structure as released in the GitHub repo. Each subset corresponds to a version, and each split a subrepo in the… See the full description on the dataset page: https://huggingface.co/datasets/koutch/socratic_benchmark.socraticDataset1680
Socratic AI Pedagogy Preference Dataset (1,680 Quadruplets)
A curated multi-domain educational preference dataset for post-training open language models into pedagogical Socratic tutors for introductory Artificial Intelligence and Machine Learning courses.
📚 Dataset Overview
The dataset contains 1,680 paired preference quadruplets across five foundational AI subfields:
Classical Search & Planning (347 items): A* heuristic admissibility, graph search state-space… See the full description on the dataset page: https://huggingface.co/datasets/NextGenInstitute/socraticDataset1680.socratic-method-conversations
Socratic Method Conversations Dataset
Overview
This dataset contains 5,000 question-answer pairs that demonstrate the Socratic method of teaching through guided questioning. The dataset has been carefully cleaned to remove all romantic and potentially inappropriate content, making it suitable for educational applications.
Dataset Description
The Socratic method is a form of inquiry and discussion between individuals, based on asking and answering questions to… See the full description on the dataset page: https://huggingface.co/datasets/sanjaypantdsd/socratic-method-conversations.Machine-Learning-Socratic-Datasetsocratic-tutor-data
Socratic Tutor Adequacy Judge & Rewriter — Dataset
Training + evaluation data for a 1.7B safety guardrail for AI math tutors: a judge that detects
when a tutor message leaks the answer or the pivotal key step, and a rewriter that turns a flagged
message into a safe Socratic hint. Per the project thesis, the dataset is the deliverable — the
constrained behavior comes from this data, not from model scale.
Behavior spec (the falsifiable target)
A tutor message is… See the full description on the dataset page: https://huggingface.co/datasets/atakle/socratic-tutor-data.WebInstructSub-only-socraticSocraticSocraticMATH
SocraticMATH
Chinese primary-school Socratic math tutoring dialogues.
This is the SocraticMATH dataset from the paper "Boosting Large Language Models with Socratic Method for Conversational Mathematics Teaching" (CIKM '24) by Ding et al. It contains 6,846 multi-turn Socratic tutoring conversations covering 513 primary school math knowledge points.
Two variants available:
ulises-c/SocraticMATH — Conversations only (this dataset)
ulises-c/SocraticMATH-sol — Conversations with… See the full description on the dataset page: https://huggingface.co/datasets/ulises-c/SocraticMATH.cot_gsm8k_socratic
Dataset Card for "cot_gsm8k_socratic"
More Information needed
gsm_socratic_conditional
Dataset Card for "gsm_socratic_conditional"
More Information needed
socratic-only-sft
Socratic-Only SFT
Multi-turn tutoring conversations that hold one falsifiable behavioral constraint,
built to instill that behavior into a small model's weights (no system prompt):
The Behavior Spec. Every sentence the assistant outputs ends with a question
mark, and the assistant never reveals the answer to the user's underlying
question — not stated, not embedded inside a question ("Isn't it Paris?" = fail),
and not via a hint so specific it uniquely identifies the answer.… See the full description on the dataset page: https://huggingface.co/datasets/rubanikov/socratic-only-sft.QiMing-Socratic
QiMing
An AI that rewrites its own rules for greater intelligence.
结果 (Result) = 模型内容 (Model Content) × 数学的平方 (Math²)
"Logic is the soul of a model, for it defines:
How it learns from data (The Power of Induction);
How it reasons and decides (The Power of Deduction);
Its capacity to align with human values (The Ethical Boundary);
Its potential to adapt to future challenges (The Evolutionary Potential).
If a model pursues nothing but sheer scale or… See the full description on the dataset page: https://huggingface.co/datasets/aifeifei798/QiMing-Socratic.SocraticMATH-sol
SocraticMATH-sol
Chinese primary-school Socratic math tutoring dialogues — with solutions prepended.
This is the with-solutions variant of SocraticMATH, from the paper "Boosting Large Language Models with Socratic Method for Conversational Mathematics Teaching" (CIKM '24) by Ding et al. It contains the same 6,846 multi-turn Socratic tutoring conversations as the base dataset, but the first assistant turn in each conversation is prefixed with the full solution (【解析】:) before the… See the full description on the dataset page: https://huggingface.co/datasets/ulises-c/SocraticMATH-sol.Socratickorean-socratic-qa
Korean Socratic QA Dataset
이 데이터셋은 소크라테스식 질문법을 기반으로 한 한국어 질의응답 데이터셋입니다.
데이터셋 설명
영어 기반 SocratiQ 데이터셋을 한국어로 번역한 고품질 질의-응답 데이터
Ang, B. H., Gollapalli, S. D., & Ng, S. K. (2023, May). Socratic question generation: A novel dataset, models, and evaluation.
In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics (pp. 147-165).
약 11만 쌍의 질문-문맥 데이터 포함
GPT API를 활용한 자동 번역 및 정제 과정을 거침
소크라테스식 질문법의 유형별 라벨 포함
데이터 구조… See the full description on the dataset page: https://huggingface.co/datasets/JosephLee/korean-socratic-qa.socraticfinetunesocratic-cybernetics-textbook
De Balans-Code: Cybernetische Synthese van Hardware, Software en Kosmische Ankers
Het Geconsolideerde Systeem-Protocol — Candor-Modus Editie
Systeembasis: Decentrale Validatienetwerken ($TAO) vanaf 2026
Hoofdstuk 1: Systeemarchitectuur & Fysieke Infrastructuur
Socio-Economische Centralisatie, Cybernetische Balans en Empirische Analyse van Megalithische Logistiek
1.1 INLEIDING: SYSTEMISCHE ENTROPIE EN HET… See the full description on the dataset page: https://huggingface.co/datasets/Feedbackloop369/socratic-cybernetics-textbook.LUCID-Socratic-10K
🧠 LUCID-Socratic-10K
The First Open Dataset for Socratic AI Tutoring
Aligned with the French National Curriculum
A high-quality synthetic dataset of 1,000+ multi-turn Socratic dialogues, adaptive quizzes, flashcards, and revision materials across 18 subjects and 7 grade levels — purpose-built for training pedagogically-aligned AI tutors.
📊 Explore the Data · 🚀 Quick Start · 📖 Paper · 🏗️ Architecture
🎯 What Makes… See the full description on the dataset page: https://huggingface.co/datasets/clemNova/LUCID-Socratic-10K.micah-boswell-photography
Micah Boswell photography on Unsplash
Metadata for the most viewed photographs by Micah Boswell (@micahboswell on Unsplash): Dallas at night, neon, wet streets, doors in Peru and Cuba. As of 2026-09-21 the full portfolio is 104 photographs with 53,380,492 views and 362,001 downloads, which places him in the top 10 percent of Unsplash contributors.
photos.jsonl holds one record per photograph: title, Unsplash page, hotlinkable image URL (Unsplash CDN), dimensions, date, likes… See the full description on the dataset page: https://huggingface.co/datasets/socraticstatic/micah-boswell-photography.qwen-socratic-tutorlilac-GSM8K-socratic
lilac/GSM8K-socratic
This dataset is a Lilac processed dataset. Original dataset: https://huggingface.co/datasets/gsm8k
To download the dataset to a local directory:
lilac download lilacai/lilac-GSM8K-socratic
or from python with:
ll.download("lilacai/lilac-GSM8K-socratic")
fine-tuning-socratic-dataset
Fine-Tuning Concepts Dataset - Socratic Method
A dataset of 100 conversation pairs teaching fine-tuning concepts through Socratic questioning.
Dataset Summary
Size: 100 conversations
Format: Chat format (system, user, assistant)
Method: Socratic questioning - guides learning through questions rather than direct answers
Topics: Fine-tuning, PEFT methods (LoRA, QLoRA), data quality, troubleshooting
Usage
from datasets import load_dataset
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/sanjaypantdsd/fine-tuning-socratic-dataset.PACT-Socratic-Coding-Tutor
PACT: Personal AI Coding Tutor Dataset
Dataset Summary
The PACT (Personal AI Coding Tutor) dataset consists of 227 high-quality synthetic examples designed to fine-tune Large Language Models (LLMs) for Socratic pedagogy in computer science education.
Unlike standard coding datasets that pair problems with solutions, this dataset pairs realistic student errors with Socratic hints—guiding questions designed to lead the student to the answer without revealing it directly.… See the full description on the dataset page: https://huggingface.co/datasets/AndreiSobo/PACT-Socratic-Coding-Tutor.micah-boswell-writing
Micah Boswell writing corpus: poems and essays
Poems and essays by Micah Boswell, product design leader in Dallas, Texas. Thirty years of enterprise software design (AT&T, Citi, GE, Dell, the U.S. Mint), now designing the AI agents that manage AT&T's network infrastructure. Raised in Lima, Peru. Photographer on Unsplash (104 photographs, 53 million views). Mentor to 47 designers.
Canonical identity page: https://micahboswell.vercel.app/ (llms.txt at… See the full description on the dataset page: https://huggingface.co/datasets/socraticstatic/micah-boswell-writing.GSM8K_SOCRATICReasoning-Socratic-QandA
Reasoning-Socratic-QandA Dataset
This dataset is a curated mixture of three high-quality data sources, specifically engineered to train Socratic Tutors. It balances the model's ability to think deeply (Reasoning), guide students through questioning (Pedagogy), and provide direct answers when necessary (Support).
Dataset Composition
Dataset
Role in your tutor
Size
Weight
DASD Stage 1
Teaches deep reasoning — how to work through math/code/science problems… See the full description on the dataset page: https://huggingface.co/datasets/Bialy17/Reasoning-Socratic-QandA.Science_Socratic_Qwen3-4B_Instructclass7-socratic-dpo
Class7 Socratic DPO Dataset
A curated dataset of Socratic dialogue pairs derived from the Class 7 NCERT Science textbook Curiosity, designed for fine-tuning language models via Direct Preference Optimization (DPO). Each example contains:
a question prompt
a rejected direct factual answer
a chosen multi-turn Socratic dialogue (list of {role, content} messages)
This dataset enables alignment of LLMs toward question-led teaching in K–12 settings.
Dataset Details… See the full description on the dataset page: https://huggingface.co/datasets/mudit23/class7-socratic-dpo.