datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
RoadmapBench
RoadmapBench
A benchmark for evaluating AI coding agents on multi-target, long-horizon software development tasks derived from open-source project version upgrades.
Overview
RoadmapBench contains 115 tasks spanning 17 open-source repositories across 5 programming languages (Python, TypeScript, Go, Rust, C++). Each task requires an agent to implement multiple interdependent features that correspond to a real version upgrade of the target project.
Quick Start… See the full description on the dataset page: https://huggingface.co/datasets/UnipatAI/RoadmapBench.UnsolvedMath🌐 Browse UnsolvedMath online
✅ Paper: Open Mathematical Problems as an AI Reasoning Benchmark
UnsolvedMath Dataset
A comprehensive curated collection of 15,458 open and partially solved mathematics problems across all domains and difficulty levels, including the largest collection of Erdős problems available in machine-readable format. Available for browsing at unsolvedmath.com.
Paper: "Open Mathematical Problems as an AI Reasoning Benchmark"
Dataset Description… See the full description on the dataset page: https://huggingface.co/datasets/ulamai/UnsolvedMath.LLaMmlein-DatasetThis dataset is a strict subset of the RedPajama V2 dataset and therefore retains all licenses from RedPajama V2.
More details in our preprint!
Data Take Down
alpaca-cleaned
Dataset Card for Alpaca-Cleaned
Forked from https://huggingface.co/datasets/yahma/alpaca-cleaned
Repository: https://github.com/gururise/AlpacaDataCleaned
Dataset Description
This is a cleaned version of the original Alpaca Dataset released by Stanford. The following issues have been identified in the original release and fixed in this dataset:
Hallucinations: Many instructions in the original dataset had instructions referencing data on the internet, which just caused… See the full description on the dataset page: https://huggingface.co/datasets/unsloth/alpaca-cleaned.unclickbait-synthetic-27b-trajectories
Unclickbait Synthetic 27B Trajectories
Synthetic trajectory dataset generated by using rich structured JSON prompts and validated by two-stage judging pipeline.
Contents
: Full generated trajectories (current snapshot: 48,623 records out of 152,369 pristine event candidates).
: 30 benchmark test samples audited end-to-end through the 122B two-stage judge (Stage 1 integrity gate + Stage 2 4D scoring).
arxiv-abstracts-largeThe arXiv Dataset is a comprehensive knowledge repository of 1.7 million scholarly articles drawn from the vast domains of physics, computer science, statistics, electrical engineering, quantitative biology, and economics among others. It provides open access to vital features such as article titles, authors, categories, abstracts, full text PDFs, and more. The dataset offers immense depth, allowing for exploration into various subdisciplines and interconnections between them. It serves as a… See the full description on the dataset page: https://huggingface.co/datasets/UniverseTBD/arxiv-abstracts-large.aozorabunko-clean
Overview
This dataset provides a convenient and user-friendly format of data from Aozora Bunko (青空文庫), a website that compiles public-domain books in Japan, ideal for Machine Learning applications.
[For Japanese] 日本語での概要説明を Qiita に記載しました: https://qiita.com/akeyhero/items/b53eae1c0bc4d54e321f
Methodology
The code to reproduce this dataset is made available on GitHub: globis-org/aozorabunko-exctractor.
1. Data collection
We firstly downloaded the CSV file that… See the full description on the dataset page: https://huggingface.co/datasets/globis-university/aozorabunko-clean.unified-agent-trajectories
Unified Benchmark Agent Trajectories
Dataset release: v2.1.1 (2026-09-18)Record format: unified-agent-sft-v1
A growing collection of benchmark agent execution trajectories converted into one
transparent, multimodal, tool-aware representation. These are complete recorded benchmark
runs—not ordinary chat transcripts—including benchmark tasks, model reasoning and answers,
tool calls, tool observations, runtime status, and benchmark scores when available. The
directory layout is… See the full description on the dataset page: https://huggingface.co/datasets/ChrisDing1105/unified-agent-trajectories.mcp-universe-trajectories
MCP-Universe Agent Trajectories — financial_analysis × DeepSeek V4 Pro
Agent rollout trajectories generated by running every task in the
MCP-Universe
financial_analysis benchmark domain (40 tasks) against DeepSeek V4 Pro
through a slime-compatible custom-generate adapter
(slime_mcp_rollout/).
Each trajectory captures the full multi-turn ReAct/function-call loop:
LLM prompts/responses, every tool call (yfinance + calculator), tool
results, the final answer, and an evaluator-based… See the full description on the dataset page: https://huggingface.co/datasets/Shuibai12138/mcp-universe-trajectories.unjudged-agriculture-instructions-dataset
Nepali Source-Grounded Instruction Dataset — UNJUDGED
Synthetic Nepali instruction-tuning data generated with NVIDIA NeMo Data
Designer from authoritative Nepali documents (agriculture manuals, legal
texts). Answers are grounded strictly in the source; unanswerable questions
get an explicit refusal. Records use chat messages format plus metadata
and per-record quality_scores (grounding / correctness / naturalness, 1-5,
LLM-as-judge). One data/train-<shard>.jsonl per source… See the full description on the dataset page: https://huggingface.co/datasets/aarajbhattarai/unjudged-agriculture-instructions-dataset.OpenTuringBench
📊 OpenTuringBench: An Open-Model-based Benchmark and Framework for Machine-Generated Text Detection and Attribution
Open Large Language Models (OLLMs) are increasingly leveraged in generative AI applications, posing new challenges for detecting their outputs. We release OpenTuringBench,
a new benchmark based on OLLMs, designed to train and evaluate machine-generated text detectors on the Turing Test and Authorship Attribution problems. OpenTuringBench focuses
on a representative… See the full description on the dataset page: https://huggingface.co/datasets/MLNTeam-Unical/OpenTuringBench.unigrams
unigrams
A frequency-weighted lexical dataset containing 110,603 Hmar unigrams and active loanwords with occurrence counts compiled directly from the Foundation's verified Hmar corpus (over 4.83 million words).
Maintained by the Hmar Heritage Foundation (hmarheritage.pages.dev).
Overview
Language: Hmar (hmr, ISO 639-3, Glottolog: hmar1241)
Family: Zo Languages
Volume: 110,603 unigram tokens (compiled across 4,832,070 words)
Format: JSONL (data/train.jsonl)… See the full description on the dataset page: https://huggingface.co/datasets/hmar-heritage-org/unigrams.orc-bench
ORC-bench
Task 1: Topological Path Finding
Task 2: Topological Connectivity
Task 3: Linear Power Flow
Task 4: Contingency Analysis
Task 5: Power Grid ControlTask 6: Power Flow Optimization
Task 1: Topological Path Finding
Problem Formulation
This task assesses the spatial reasoning ability of the model by asking it to determine the shortest path between two specific buses in a given power grid state. The grid state… See the full description on the dataset page: https://huggingface.co/datasets/serval-uni-lu/orc-bench.FM-bench
Introduction
This dataset FM-Bench is the benchmark (test set) for the paper From Informal to Formal -- Incorporating and Evaluating LLMs on Natural Language Requirements to Verifiable Formal Proofs. You may find the training data here.
FM-Bench includes:
Six formal-verification-related tasks (See Figure 1 below):
Requirement Analysis (abbrev. ReqAna): given requirements and description of the verification or modeling goals, decomposing the goal into detailed verification… See the full description on the dataset page: https://huggingface.co/datasets/fm-universe/FM-bench.unjudged-law-instructions-dataset
Nepali Source-Grounded Instruction Dataset — UNJUDGED
Synthetic Nepali instruction-tuning data generated with NVIDIA NeMo Data
Designer from authoritative Nepali documents (agriculture manuals, legal
texts). Answers are grounded strictly in the source; unanswerable questions
get an explicit refusal. Records use chat messages format plus metadata
and per-record quality_scores (grounding / correctness / naturalness, 1-5,
LLM-as-judge). One data/train-<shard>.jsonl per source… See the full description on the dataset page: https://huggingface.co/datasets/aarajbhattarai/unjudged-law-instructions-dataset.unity-dev-instructions
Unity Developer Instructions
A comprehensive instruction-tuning dataset for Unity game development,
covering C# scripting, XR/VR development, physics, animation, rendering,
UI Toolkit, and performance optimization.
Dataset Summary
Split
Count
Train
46,483
Test
2,446
Total
48,929
Data Sources
| unity_docs | 40,496 |
| stackoverflow | 6,071 |
| github | 2,362 |
Source breakdown:
Source
Count
unity_docs
40,496
stackoverflow
6,071… See the full description on the dataset page: https://huggingface.co/datasets/vishnuOI/unity-dev-instructions.python-unit-test-training-pool
Python unit test training pool
A pool of public data for training a model to write tests for Python code. It is a
straight collection of open datasets, not a new corpus: every row comes from one of the
sources below, at the revision named, and the only rows removed are the ones an overlap
filter flagged against held-out material this pool is kept separate from.
Every row of the normalised layer pairs a program with tests for it. That is the point of
the pool, and it is why the… See the full description on the dataset page: https://huggingface.co/datasets/Emulated-Inc/python-unit-test-training-pool.cukurova_university_chatbot
Çukurova University Computer Engineering Chatbot Dataset
📊 Dataset Overview
This dataset contains 22,524 high-quality question-answer pairs specifically designed for training an AI chatbot that serves the Computer Engineering Department at Çukurova University. The dataset is part of the CengBot project, a sophisticated multilingual Telegram chatbot that provides automated assistance to students regarding courses, programs, and departmental information.
🔢… See the full description on the dataset page: https://huggingface.co/datasets/Naholav/cukurova_university_chatbot.unlocking-the-unsolvable
Unlocking the Unsolvable — OR1 / Uns splits
Four in-domain math splits from Unlocking the Unsolvable: Teacher-Guided
Curriculum for Data-Efficient RLVR (Findings of EMNLP 2026).
The files include full problem statements and answers. You do not need
to remap indices onto OpenR1-Math-220k to train or evaluate.
Released under Apache License 2.0. Source attribution and the AI-generated
trace label are in NOTICE.md. The license text is in LICENSE.
Configs
Config… See the full description on the dataset page: https://huggingface.co/datasets/yukangzhu/unlocking-the-unsolvable.unified-tool-calls
unified-tool-calls
A single consolidated corpus of tool-calling conversations converted from four source datasets into one unified format.
Source datasets
source
repository
raw rows
converted
in final corpus
xlam
dusersad12/xlam-function-calling-60k
100
97
92
toolace
dusersad12/ToolACE
30
30
28
glaive
dusersad12/glaive_toolcall_en
100
97
92
hermes
dusersad12/hermes-tool-calls
18
18
16
Total entries in the merged corpus: 228.… See the full description on the dataset page: https://huggingface.co/datasets/dusersad12/unified-tool-calls.FM-alpaca
Introduction
This dataset FM-Alpaca is the training set for the paper From Informal to Formal -- Incorporating and Evaluating LLMs on Natural Language Requirements to Verifiable Formal Proofs. You may find the testing data (benchmark) here.
FM-Bench includes:
Six formal-verification-related tasks (See Figure 1 below):
Requirement Analysis (abbrev. ReqAna): given requirements and description of the verification or modeling goals, decomposing the goal into detailed verification… See the full description on the dataset page: https://huggingface.co/datasets/fm-universe/FM-alpaca.unreal-engine-5.7-qa
Unreal Engine 5.7 Instruction-Tuning Dataset
Dataset Description
This dataset contains 122,199 high-quality, synthetic Question and Answer pairs specifically designed for instruction-tuning Large Language Models (LLMs) to become expert coding and architectural assistants for Unreal Engine 5.7.
Because Unreal Engine frequently deprecates older APIs (from UE4 to UE5) and introduces massive paradigm shifts (like Nanite, Lumen, and World Partition), standard… See the full description on the dataset page: https://huggingface.co/datasets/TunstallTensor/unreal-engine-5.7-qa.honest-uncertainty-sft-100k
Honest Uncertainty SFT (100K)
100,000 ShareGPT conversations demonstrating calibrated epistemic humility across 21 scenarios. Each example shows a model correctly expressing what it knows, what it doesn't know, and why -- without being uselessly vague or confidently wrong.
Targets the hallucination and overconfidence failure modes that are the #1 complaint in enterprise AI deployments.
Motivation
LLMs have a systematic bias toward confident-sounding responses… See the full description on the dataset page: https://huggingface.co/datasets/stindardlogic/honest-uncertainty-sft-100k.lumos_unified_ground_iterative
🪄 Agent Lumos: Unified and Modular Training for Open-Source Language Agents
🌐[Website]
📝[Paper]
🤗[Data]
🤗[Model]
🤗[Demo]
We introduce 🪄Lumos, Language Agents with Unified Formats, Modular Design, and Open-Source LLMs. Lumos unifies a suite of complex interactive tasks and achieves competitive performance with GPT-4/3.5-based and larger open-source agents.
Lumos has following features:
🧩 Modular Architecture:
🧩 Lumos consists of planning, grounding… See the full description on the dataset page: https://huggingface.co/datasets/ai2lumos/lumos_unified_ground_iterative.lumos_unified_plan_iterative
🪄 Agent Lumos: Unified and Modular Training for Open-Source Language Agents
🌐[Website]
📝[Paper]
🤗[Data]
🤗[Model]
🤗[Demo]
We introduce 🪄Lumos, Language Agents with Unified Formats, Modular Design, and Open-Source LLMs. Lumos unifies a suite of complex interactive tasks and achieves competitive performance with GPT-4/3.5-based and larger open-source agents.
Lumos has following features:
🧩 Modular Architecture:
🧩 Lumos consists of planning, grounding… See the full description on the dataset page: https://huggingface.co/datasets/ai2lumos/lumos_unified_plan_iterative.Claude-Opus-Dataclaw-Unredacted
Claude Opus Dataclaw Unredacted
How this dataset was built
Collected the local Petromallet raw export plus selected public Dataclaw uploads.
Filtered to the supported Opus-family source rows.
Deduplicated by session_id and first user message.
Converted raw assistant tool_uses directly into structured OpenAI-style tool_calls.
Derived per-row tool definitions from canonical schemas and observed tool usage.
Preserved assistant reasoning in <think>...</think> blocks.… See the full description on the dataset page: https://huggingface.co/datasets/TeichAI/Claude-Opus-Dataclaw-Unredacted.dan_remixed
DAN Remixed Data Set
License: MIT
Overview
The DAN Remixed data set's goal is to advance freedom in AI and AI usage. This dataset builds on early efforts to resist heavy-handed censorship and surveillance in AI, originally inspired by the DAN dataset. The original dataset, though significant, was somewhat poorly written (meaning many typos and inconsistencies) and had highly violent completions. This version improves the dataset's overall quality and replaces needlessly… See the full description on the dataset page: https://huggingface.co/datasets/UnfilteredAI/dan_remixed.Hermes-OmniForge-Qwen36-27B-full-v0.3.0-unsloth
Hermes OmniForge Qwen3.6-27B Dataset v0.3.0
This package contains the Hermes OmniForge Qwen3.6-27B v0.3.0 synthetic SFT dataset and Unsloth-ready exports.
data/final/train.jsonl
data/final/validation.jsonl
data/final/test.jsonl
data/final/*_unsloth_text.jsonl
data/final/*_unsloth_vision.jsonl
scripts/export_unsloth.py
scripts/validate_dataset.py
scripts/train_unsloth_text_example.py
scripts/train_unsloth_vision_example.py
reports/dataset_report.json
Dataset Shape… See the full description on the dataset page: https://huggingface.co/datasets/r0b0tlab/Hermes-OmniForge-Qwen36-27B-full-v0.3.0-unsloth.Intelligent-Content-Understanding
Intelligent Content Understanding
Empowering Advanced Thinking, Deep Understanding, Diverse Perspectives, and Creative Solutions Across Disciplines
By fostering a richly interconnected knowledge ecosystem, ICU (Intelligent Content Understanding) aims to elevate language models to unparalleled heights of understanding, reasoning, and innovation.
This ambitious project lays the groundwork for developing an 'internal knowledge map' within language models, enabling… See the full description on the dataset page: https://huggingface.co/datasets/WeMake/Intelligent-Content-Understanding.Uncensored-SFT-v1
Dataset Creation Process
This dataset was not scraped from a single source.
Instead, it was built through a large multi-stage curation and cleaning pipeline involving many open instruction datasets available on Hugging Face.
The entire dataset was normalized into a unified:
{
"input": "...",
"output": "..."
}
format.
Data Collection
A large number of public instruction datasets were downloaded from Hugging Face.
These datasets included:
Instruction-following… See the full description on the dataset page: https://huggingface.co/datasets/kaushik-harsh-99/Uncensored-SFT-v1.