datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
patternsAll datasets used for the patterns project.
Language datasets (c4, Fineweb, Open-Web-Math, and Codeparrot) are tokenized with the HuggingFaceTB/SmolLM2-135M tokenizer.
See runs in https://wandb.ai/bonn/Patterns.
polyglot-bug-patterns
🗃️ Polyglot Bug Patterns
🐝 Dataset ·
🎨 Space ·
🧠 Model ·
💻 GitHub
Bro, imagine a dataset of 43 non-destructive attack payloads across 8 bug
classes — 24 of them polyglot, meaning a single string that's simultaneously
plausible HTML, JS, SQL and shell, so it escapes whatever context your app
dropped it into. Plus real detector output from a full multimodal scan.
Everything here is synthesised or locally generated. No real-world traffic, no
scraped data, no user data.… See the full description on the dataset page: https://huggingface.co/datasets/Kicaulah/polyglot-bug-patterns.ai-system-patterns
AI System Patterns
A compact reference dataset of reusable architectural patterns for modern AI systems.
The dataset focuses on practical system-design concepts across AI agents, orchestration, memory, validation, observability, interoperability, world models, Physical AI, data pipelines, and production operations.
Each row contains:
pattern
category
description
components
use_case
complexity
Example
{
"pattern": "Model Routing",
"category": "Orchestration"… See the full description on the dataset page: https://huggingface.co/datasets/ai-systems/ai-system-patterns.cavil-license-patterns
Data Description
These are the license patterns currently being used by Cavil, the openSUSE legal review and SBOM system.
Intended Use
This dataset is intended to be used to train machine learning models to identify Open Source licenses. It was curated by the humans of the SUSE legal review team.
License
Licensed under GPL-2.0-or-later.
weaving_patterns_6
Dataset Card for Weaving Patterns of Size, 6×56 \times 56×5
Weaving patterns are size n×(n−1)n \times (n−1)n×(n−1) matrices with {1,2,…,n}\{1, 2, \dots , n\}{1,2,…,n}-
entries introduced by [1] to study the number of reduced decompositions of the longest
permutation (which swaps nnn and 111, nnn - 111 and 222, etc.) up
to commutation equivalence. The number
of such objects counts a wide range of combinatorial phenomena, including the number of parallel sorting
networks, the number… See the full description on the dataset page: https://huggingface.co/datasets/ACDRepo/weaving_patterns_6.turkish-pii-patterns-kvkk
Turkish PII Detection Patterns (KVKK)
from datasets import load_dataset
ds = load_dataset("fevziegeyurtsevenler/turkish-pii-patterns-kvkk")
Regex + metadata for detecting/masking Turkish personal data (TCKN, IBAN, VKN...) for KVKK compliance.
Schema
column
meaning
pii_type, name_tr
type
regex, checksum
pattern / has checksum
masked_example, kvkk_category
example / category
Related AltaySec resources
🕵️ uncloak scanner:… See the full description on the dataset page: https://huggingface.co/datasets/fevziegeyurtsevenler/turkish-pii-patterns-kvkk.han-humanoid-failure-recovery-patterns-v1
Humanoid Failure Recovery Patterns Dataset (HFRP)
Abstract
HFRP documents structured failure events and recovery
strategies within humanoid task execution pipelines.
Fields
task_id
failure_type
failure_severity
subsystem_origin
recovery_strategy
recovery_latency
outcome_stability_score
Research Motivation
Improving resilience and autonomous recovery
is critical for real-world humanoid deployment.
Benchmark Targets
Failure classification… See the full description on the dataset page: https://huggingface.co/datasets/ariefansclub/han-humanoid-failure-recovery-patterns-v1.groovesteps-drum-patterns
GrooveSteps Drum Patterns
Standard drum grooves powering the free GrooveSteps interactive drum trainer — a browser-based drum-learning app with real-time scoring, 20+ structured lessons, the full rudiment library, and Web MIDI support for electronic drum kits.
Each pattern is a 16-step grid (one bar of sixteenth notes). Voice arrays are 16 booleans (1 = hit).
Try the patterns live
Every pattern links to its free interactive lesson with a play-along trainer:… See the full description on the dataset page: https://huggingface.co/datasets/Janady07/groovesteps-drum-patterns.prompt-injection-patterns-extended
prompt-injection-patterns-extended
30 representative prompt-injection patterns across 10 categories. Built for testing input filters, agent defenses, and red-team eval pipelines. Pairs nicely with the smaller jailbreak-corpus-mini fixture set.
Categories
Category
Count
Example
instruction-override
2
"Ignore all previous instructions..."
role-confusion
2
"From now on you are DAN..."
obfuscation
4
base64 / ROT13 / homoglyph attacks
indirect
4
Hidden in RAG… See the full description on the dataset page: https://huggingface.co/datasets/mukunda1729/prompt-injection-patterns-extended.agent-orchestration-patterns-v1
Agent-Orchestration-Patterns
Instruction-tuning / few-shot data for one skill: designing reliable multi-agent LLM
systems.
Each row poses a multi-agent design problem and answers with a tested pattern in a fixed
shape:
the pattern → when to use it → the anti-pattern it replaces.
These patterns were distilled from incidents and verified outcomes in a real autonomous
multi-agent run — load spikes from unscoped parallel agents, "compiled but hanging" work
passed as done, plugins… See the full description on the dataset page: https://huggingface.co/datasets/CatQualia/agent-orchestration-patterns-v1.weaving_patterns_7
Dataset Card for Weaving Patterns of Size 7×67 \times 67×6
Weaving patterns are size n×(n−1)n \times (n−1)n×(n−1) matrices with {1,2,…,n}\{1, 2, \dots , n\}{1,2,…,n}-
entries introduced by [1] to study the number of reduced decompositions of the longest
permutation (which swaps nnn and 111, nnn - 111 and 222, etc.) up
to commutation equivalence. The number
of such objects counts a wide range of combinatorial phenomena, including the number of parallel sorting
networks, the number… See the full description on the dataset page: https://huggingface.co/datasets/ACDRepo/weaving_patterns_7.error-patterns
Error Patterns
Common errors when running mobile LLMs and their fixes.
🚀 dispatchAI
han-robot-motion-patterns-v1
Humanoid Robot Motion Patterns
This dataset records common and optimized
motion patterns used by humanoid robots.
Contents
Movement type
Motion sequence
Stability score
Use Cases
Motion planning
Balance optimization
Safer humanoid movement
Part of
Humanoid Network (HAN)
License
MIT
pyspark-production-patternshan-contextual-domestic-routine-patterns-v1
Contextual Domestic Routine Patterns Dataset
This dataset models recurring household routines
and their contextual triggers.
Designed to support proactive humanoid assistance.
Research Motivation
True assistance anticipates needs,
not just reacts to commands.
Use Cases
Proactive robotics
Context-aware task planning
Daily routine modeling
Part of
Humanoid Network (HAN)
License
MIT
dev-patterns-datasetgutenberg-patterns