datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Unified_Agent_Framework
A Unified Framework for the Evaluation of LLM Agentic Capabilities
This repository contains the dataset (Benchmark, Toolkit, and Environment assets) for the paper A Unified Framework for the Evaluation of LLM Agentic Capabilities.
The official code and agent execution sandbox can be found on GitHub: whfeLingYu/A-Unified-Framework-for-the-Evaluation-of-LLM-Agentic-Capabilities.
Dataset Description
The dataset integrates diverse agent benchmarks into a standardized… See the full description on the dataset page: https://huggingface.co/datasets/whfeLingYu/Unified_Agent_Framework.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.unified-toolcalls-canonical
Unified Tool-Calling Corpus — Canonicalized Output
Publish-ready conversion of two pinned Hugging Face dataset revisions into the single
schema defined in docs/unified_format.md, with repeated
records normalized by an explicit canonicalization rule and every surviving record
kept faithful to its source row.
Records in (source rows)
65,000
Records published (canonical survivors)
64,622
Duplicates collapsed
378 (343 duplicate groups)
Records mutated during… See the full description on the dataset page: https://huggingface.co/datasets/dongbobo/unified-toolcalls-canonical.portuguese-unified-pronunciation-lexicon
Portuguese Unified Pronunciation Lexicon
A flat, single-row-per-pronunciation dataset merging Portuguese IPA transcriptions from three authoritative sources. Each row is a word × region × POS tuple with both broad phonemic (ipa_broad) and narrow phonetic (ipa_narrow) transcriptions normalized across sources.
Source
Words
Convention
Description
Infopédia (Porto Editora)
102,685
Broad phonemic
European Portuguese dictionary IPA
Wiktionary (pt.wiktionary.org)
15,720… See the full description on the dataset page: https://huggingface.co/datasets/TigreGotico/portuguese-unified-pronunciation-lexicon.unified-reasoning-dataset
Unified Reasoning Dataset
A 94,860-row English SFT collection that normalizes four synthetic reasoning and instruction datasets into one consistent schema.
Quick start
from datasets import load_dataset
dataset = load_dataset(
"j0no12/unified-reasoning-dataset",
split="train",
)
print(dataset.column_names)
# ['thinking', 'instruction', 'response', 'source']
print(dataset[0])
Dataset summary
Property
Value
Split
train only
Rows… See the full description on the dataset page: https://huggingface.co/datasets/j0no12/unified-reasoning-dataset.code-verl-unified
Unified Code VERL Dataset
Overview
This dataset aggregates seven code-reasoning collections into a single VERL-formatted repository containing approximately 958,539 unique problems. The compilation prioritizes consistent extra_info structure across all source materials for seamless compatibility with VERL training frameworks.
Dataset Composition
Seven distinct splits comprise the collection:
Split
Problems
Percentage
Format
kodcode_v1_verl
434,876… See the full description on the dataset page: https://huggingface.co/datasets/sungyub/code-verl-unified.unified-vulnerability-intelligence-dataset
Unified Vulnerability Intelligence Dataset (UVID) v3.0 — Cyber Security Knowledge Graph
UVID is a structured cyber security knowledge graph that unifies multiple
vulnerability classification frameworks into a single knowledge base. Each of the
250 records describes one application/software security vulnerability and links
it — where authoritative data exists — across CWE, CAPEC, MITRE ATT&CK, CVSS,
14 OWASP projects, secure-fix intelligence, detection surfaces, programming… See the full description on the dataset page: https://huggingface.co/datasets/ismailtasdelen/unified-vulnerability-intelligence-dataset.moe-unified-dataset-sota
moe-unified-dataset-sota
A unified dataset for training Mixture of Experts (MoE) models, combining multiple high-quality sources.
Dataset Statistics
Total Examples: 2,186,763
Train Split: 2,077,424
Test Split: 109,339
Sources
NousResearch/Hermes-3-Dataset - General instruction following, math, coding (~950k examples)
Salesforce/xlam-function-calling-60k - Function/tool calling (60k examples)
MegaScience/TextbookReasoning - Academic Q&A (~650k examples)… See the full description on the dataset page: https://huggingface.co/datasets/Yxanul/moe-unified-dataset-sota.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.math-verl-unified
Math-VERL Unified Collection
📊 Dataset Summary
A unified collection of 9 high-quality mathematical reasoning datasets totaling 2,269,166 deduplicated problems, all converted to VERL format for reinforcement learning applications. This collection combines diverse mathematical content from competition-level problems to advanced reasoning tasks.
Key Features:
2.27M deduplicated samples from 9 curated sources
Inter-dataset deduplication applied (v3.0) - 12.7%… See the full description on the dataset page: https://huggingface.co/datasets/sungyub/math-verl-unified.tunisian-derja-unified-raw-corpusTunisian Derja Unified Raw Corpus
Dataset Description
Repository: hamzabouajila/tunisian-derja-unified-raw-corpus
Paper: Not yet published; dataset card serves as primary documentation
Point of Contact: Hamza Bouajila
License: CC-BY-SA-4.0
Dataset Summary
The Tunisian Derja Unified Raw Corpus is a comprehensive collection of ~802,659 text examples in Tunisian Arabic (Derja), a low-resource dialect of Arabic widely spoken in Tunisia. This raw corpus aggregates data from multiple sources… See the full description on the dataset page: https://huggingface.co/datasets/hamzabouajila/tunisian-derja-unified-raw-corpus.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.unified-sft-dataset
Loading
from datasets import load_dataset
ds = load_dataset("himalaya-ai/unified-sft-dataset")
qa-verl-unified
Unified QA VERL Dataset
A unified collection of 5 high-quality question-answering and reasoning datasets in VERL format, deduplicated and optimized for reinforcement learning training.
Dataset Summary
This dataset combines 5 diverse QA and reasoning datasets into a single unified collection:
Total Problems: 86,379 unique problems (after 0.00% deduplication)
Original Size: 0 problems (before deduplication)
Format: VERL (Volcano Engine Reinforcement Learning)
Language:… See the full description on the dataset page: https://huggingface.co/datasets/sungyub/qa-verl-unified.UnifiedInstruct-709k
Mixed Alpaca Math Code Science Instruct
A mixed Alpaca-style instruction dataset containing math, code, science, commonsense, and general instruction examples.
The dataset is intended for supervised fine-tuning and instruction-tuning experiments, especially for small language models for educational purpose. :)
Dataset Splits
Split
Examples
Train
600,000
Validation
54,747
Test
54,748
Sources and Credits
This dataset was created by… See the full description on the dataset page: https://huggingface.co/datasets/srmty/UnifiedInstruct-709k.unified-tool-call-corpus
Unified Tool-Call Corpus
A small, cleaned corpus of tool-calling (function-calling) conversations in a single
unified schema, built for fine-tuning small function-calling models.
It merges the leading chunk of three public tool-calling dataset exports
(ToolACE-style, glaive-style and XLAM-style), each of which encodes
messages, tool calls and tool definitions differently, into one format.
Format
Each JSON Lines record of unified_tool_call.jsonl is one… See the full description on the dataset page: https://huggingface.co/datasets/dusersad12/unified-tool-call-corpus.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.unified-sft-dataset
Loading
from datasets import load_dataset
ds = load_dataset("himalaya-ai/unified-sft-dataset")
Nogai-Unified-Corpus-v1
Nogai Unified Corpus (NUC) v1
Monolingual text in Nogai (Kipchak Turkic, Cyrillic script) for continued pre-training of language models. It was built for NogaiLLM.
Size
Rows (sentences / short paragraphs)
163,531: 155,354 train, 8,177 validation (95 / 5)
Words
2.35 M
Characters
18.1 M
UTF-8 text
33.4 MB (files: 35.6 MB)
Tokens (Qwen2.5 tokenizer)
9.8 M
Format
JSONL, one {"text": ...} per row
Sources
Newspapers: digitised… See the full description on the dataset page: https://huggingface.co/datasets/ansarzeinulla/Nogai-Unified-Corpus-v1.unified_medical_scenario_benchmark
Unified Medical Scenario Benchmark
This dataset unifies three medical AI evaluation sources into a single CUBE-oriented tabular schema for future prompt/model/tool routing experiments:
MMLU-Medical from cais/mmlu, using medical subject configs only.
HealthBench from openai/healthbench, using the full 5,000-example set plus the 1,000-example hard subset marker.
MedAgentBench from bdanko/medagentbench, converted from the local medagentbench_cube implementation.
Splits… See the full description on the dataset page: https://huggingface.co/datasets/bdanko/unified_medical_scenario_benchmark.mizushi-orpo-unified
mizushi-orpo-unified
Preference pairs for ORPO training of a small language model that draws
styled vector glyphs as SVG paths. Every pair is::
prompt system + user, asking for a styled path of one character
chosen a real drawing of that character, marker-wrapped
rejected a NEGATIVE for that character, drawn by a model or damaged
What makes these negatives interesting
They are not random. Each rejected_kind is a different, measured
failure of a real… See the full description on the dataset page: https://huggingface.co/datasets/shibadogcap/mizushi-orpo-unified.aultra-unified-training-data
AUltra Unified Training Data
This dataset package contains the reconstructed chat-format training data used for the AUltra Unified defensive cybersecurity and code-assistant fine-tune.
The dataset was reconstructed from the original preparation scripts, deterministic seeds, local Hugging Face cache, and the same public upstream dataset. The reconstructed split sizes match the documented training run.
Transparency Notice
This dataset is an experimental, partially… See the full description on the dataset page: https://huggingface.co/datasets/Anes-03/aultra-unified-training-data.CFR-Title-10-SEC-167B-Unified-Combatant-Command-for-Cyber-Operations
10 U.S.C. § 167b Unified Combatant Command for Cyber Operations
Maintainer: Terry Eppler
Owner: US Federal Government
Dataset Summary
This dataset contains 250 document-grounded question-and-answer records based on 10 U.S.C. § 167b, Unified Combatant Command for Cyber Operations, together with the amendment notes, statutory notes, related implementation provisions, and executive memorandum reproduced with the section.
The source establishes the United… See the full description on the dataset page: https://huggingface.co/datasets/leeroy-jankins/CFR-Title-10-SEC-167B-Unified-Combatant-Command-for-Cyber-Operations.unifiedor-100k
UnifiedOR-100K
Unified Operations Research foundation dataset combining eight heterogeneous OR benchmarks into a single schema with multi-layer representations.
Source Datasets
Source
Hub Reference
OR Layer
FrontierCO
alirezaaminzadeh/frontierco-instance-features
Combinatorial optimization + solver performance
Text2Opt-Bench
alirezaaminzadeh/opticoder-binding-cases
NL → MILP binding
OptMATH
nvidia/OptiMATH-Train
Math word problems
Learn2Zinc… See the full description on the dataset page: https://huggingface.co/datasets/alirezaaminzadeh/unifiedor-100k.aksara-unified-datasets-v1
AksaraLLM/aksara-unified-datasets-v1
Unified AksaraLLM datasets generated from TPU output.
Files
aksara_sft_base.jsonl
aksara_sft_synthetic.jsonl
aksara_sft_final.jsonl
aksara_sigap.jsonl
aksara_dpo.jsonl
aksara_pretrain_seed.jsonl
dataset_manifest.json
Counts
{
"sft_base": 44793,
"sft_synthetic": 583,
"sft_final": 45376,
"sigap": 135,
"dpo": 6000,
"pretrain_seed": 45390
}
Final Sources
{
"aksarallm_v3": 1649,
"sigap_clean_v2": 130… See the full description on the dataset page: https://huggingface.co/datasets/AksaraLLM/aksara-unified-datasets-v1.if-verl-unified
Unified IF VERL Dataset
A unified collection of 2 high-quality instruction-following (IF) evaluation datasets in VERL format, deduplicated and optimized for reinforcement learning training.
Dataset Summary
This dataset combines 2 instruction-following evaluation datasets into a single unified collection:
Total Problems: 110,342 unique problems (after 0.00% deduplication)
Original Size: 110,346 problems (before deduplication)
Constraint Types: 79 unique constraint types… See the full description on the dataset page: https://huggingface.co/datasets/sungyub/if-verl-unified.New_unified
Qwen Unified Security Dataset
A unified, multi-task dataset for training and evaluating Large Language Models on code security tasks:
Vulnerability Detection (Binary classification)
CWE Classification (Multi-class CWE identification)
Compiled from TitanVul and REEF datasets (excluding C# and BenchVul).
simson-unified-knowledge-graph
🧠 Simson Unified Knowledge Graph
173 Nodes × 348 Edges – der Klebstoff zwischen allen Simson-Datasets.
Was das ist
Ein maschinenlesbarer Graph, der alle 6 Datasets miteinander verknüpft:
Dataset
Status
Nodes
racing-planet-simson-traces
Diagnose-Traces
15
simson-forum-qa-pairs
Forum-Wissen
30
simson-repair-manual
Technische Daten
14
racing-planet-product-catalog
Teilekatalog
37
simson-youtube-tutorials
Video-Tutorials
20… See the full description on the dataset page: https://huggingface.co/datasets/jmp1987/simson-unified-knowledge-graph.ccisd-unified-master-2024
CCISD Unified School Master (2024)
School-level records for Clear Creek Independent School District (Texas), compiled from
the district's public school pages and Texas Education Agency accountability reports.
Covers 39 schools with principal names, contact details, enrollment, and accountability
ratings.
Loading
from datasets import load_dataset
ds = load_dataset("robworks-software/ccisd-unified-master-2024")
all_schools = ds["full"] # all 39 schools… See the full description on the dataset page: https://huggingface.co/datasets/robworks-software/ccisd-unified-master-2024.kurdish-unified-corpus
Unified Kurdish Corpus
Dataset Description
This dataset aggregates 758,166 Kurdish text samples from multiple high-quality sources. All text has been preprocessed using asosoft.
Languages
Central Kurdish (ckb)
Kurdish (ku)
Dataset Structure
Columns
text: Preprocessed text content (asosoft applied)
base_dataset: Source dataset name
url: Source URL (NULL if not available)
word_count: Number of words (space-separated)
character_count:… See the full description on the dataset page: https://huggingface.co/datasets/shiima/kurdish-unified-corpus.