datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
ATO-Australian-Tax-Rulings-and-Guidance
ATO Rulings & Guidance — Australian Tax Law, Structured for AI
67,000+ Australian Taxation Office documents as RAG-ready NDJSON/CSV — Edited Private Advice, public rulings and determinations, ATO Interpretative Decisions, practical compliance guidelines, taxpayer alerts, decision impact statements, practice statements and legislative instruments. Every document parsed into structured, typed fields for legal RAG, LLM fine-tuning, and tax research automation.
Machine-readable… See the full description on the dataset page: https://huggingface.co/datasets/simplelex/ATO-Australian-Tax-Rulings-and-Guidance.BenchMAX_Rule-based
Dataset Sources
Paper: BenchMAX: A Comprehensive Multilingual Evaluation Suite for Large Language Models
Link: https://huggingface.co/papers/2502.07346
Repository: https://github.com/CONE-MT/BenchMAX
Dataset Description
BenchMAX_Rule-based is a dataset of BenchMAX, sourcing from IFEval, which is a rule-based benchmark for evaluating the instruction following capabilities in multilingual scenarios.
We extend the original dataset to 16 non-English languages by first… See the full description on the dataset page: https://huggingface.co/datasets/LLaMAX/BenchMAX_Rule-based.ruler-full
RULER Benchmark (Full)
Complete RULER benchmark dataset with all 13 tasks across 6 context lengths (4K to 128K tokens).
Overview
Metric
Value
Total Samples
78,000 (39,000 per variant)
Tasks
13
Context Lengths
4K, 8K, 16K, 32K, 64K, 128K
Samples per Config
500
Variants
memwrap, plain
Tasks
Retrieval (NIAH - Needle in a Haystack)
niah_single_1, niah_single_2, niah_single_3 - Single needle variants
niah_multikey_1… See the full description on the dataset page: https://huggingface.co/datasets/tonychenxyz/ruler-full.rule34xyz
Dataset Card for rule34.xyz
Dataset Summary
This dataset contains information about image files from rule34.xyz, a booru-style imageboard. The dataset includes metadata for 590,983 image files, including URLs, tags, file information, and like counts. The actual image files are stored in zip archives, with each archive containing 1000 image files. The data collection cutoff for this dataset is end of August/early September 2024.
Languages
The dataset metadata is… See the full description on the dataset page: https://huggingface.co/datasets/nyuuzyou/rule34xyz.swiss_rulings
Dataset Card for Swiss Rulings
Dataset Summary
SwissRulings is a multilingual, diachronic dataset of 637K Swiss Federal Supreme Court (FSCS) cases. This dataset can be used to pretrain language models on Swiss legal data.
Supported Tasks and Leaderboards
Languages
Switzerland has four official languages with three languages German, French and Italian being represenated. The decisions are written by the judges and clerks in the language of the… See the full description on the dataset page: https://huggingface.co/datasets/rcds/swiss_rulings.RuleWeaverRuleWeaver
This directory contains the public evaluation set for the paper RuleWeaver:
Benchmarking Rule-Centered Scenario Reasoning for Large Language Models.
To evaluate models on this benchmark, use the evaluation code provided in the
RuleWeaver GitHub repository.
File
Contents
scenario_qa.jsonl
96 scenario-based QA cases: 48 same-source and 48 cross-source.
rules.jsonl
The organized pool of 200 root rules, 50 per source dataset, with four final variants per… See the full description on the dataset page: https://huggingface.co/datasets/SharkSpicy/RuleWeaver.rule34lol-images-part2
Dataset Card for rule34lol-images-part2
Dataset Summary
This dataset contains information about image files from rule34.lol, a booru-style imageboard. The dataset includes metadata for 77,000 image files, including URLs, tags, file information, and like counts. The actual image files are stored in zip archives, with each archive containing 1000 image files (except the last archive). This is Part 2 of 2 for the complete rule34lol-images dataset. Part 1 can be found here.… See the full description on the dataset page: https://huggingface.co/datasets/nyuuzyou/rule34lol-images-part2.lm-eval-ruler-results-private-32K
Dataset Card for Evaluation run of elichen3051/Llama-3.1-8B-GGUF
Dataset automatically created during the evaluation run of model elichen3051/Llama-3.1-8B-GGUF
The dataset is composed of 12 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An… See the full description on the dataset page: https://huggingface.co/datasets/elichen-skymizer/lm-eval-ruler-results-private-32K.amc12_22-24canonical-dataset
canonical-dataset
A parallel, multilingual dataset on rule-following
Languages
en — English
am — Amharic
de — German
hi — Hindi
ig — Igbo
it — Italian
ko — Korean
ru — Russian
sw — Swahili
ta — Tamil
tr — Turkish
ur — Urdu
yo — Yoruba
Loading
from datasets import load_dataset
en = load_dataset("canonical-dataset", "en", split="test")
yo = load_dataset("canonical-dataset", "yo", split="test")
cross_rulings_hts_dataset_for_tariffs
CROSS Rulings HTS Dataset for Tariff Classification
Maintained by Flexify.AI Inc. as part of the ATLAS trade intelligence research program.
Paper: ATLAS: Benchmarking and Adapting LLMs for Global Trade via Harmonized Tariff Code Classification
Project Page: https://tariffpro.flexify.ai/
This dataset is constructed from the U.S. Customs and Border Protection (CBP) Rulings Online Search System (CROSS).It contains rulings where importers sought clarification on the correct Harmonized… See the full description on the dataset page: https://huggingface.co/datasets/Dayanand314Krishna/cross_rulings_hts_dataset_for_tariffs.ru-law-retrieval
RuLawRetrieval: поиск статей законов РФ по вопросам
Бенчмарк поиска (retrieval) по кодексам РФ в формате MTEB (corpus / queries / qrels) со сплитами train / dev / test и подмножеством test с повторной разметкой и множественной релевантностью (golden, размечено ИИ-агентом). Сделан в проекте ru-law-retrieval, где на нём сравниваются 12 готовых эмбеддеров, BM25 и дообученная multilingual-e5-small.
English summary: a Russian legal retrieval benchmark (MTEB format). Corpus: 3,786… See the full description on the dataset page: https://huggingface.co/datasets/alekseevpavel04/ru-law-retrieval.multimodality-poc-llama31-ruler16k
Multimodality PoC corpus — Llama-3.1-8B-Instruct on RULER-16K
Raw pre-RoPE query and hidden-state tensors captured during prefill, used
to study whether the per-(layer, kv_head) query distribution is unimodal
Gaussian (the assumption underpinning Expected Attention's MGF closed-form
in kvpress).
What's in here
65 .npz files, one per (RULER task, prompt_index) pair (13 tasks × 5
prompts).
Each file (~414 MB) contains:
field
dtype
shape
meaning
hidden
float16… See the full description on the dataset page: https://huggingface.co/datasets/June30916/multimodality-poc-llama31-ruler16k.cybersecurity-rules
Cybersecurity Detection Rules Dataset
This dataset contains a collection of 950 detection rules from official SIGMA, YARA, and Suricata repositories. Knowledge distillation was applied to generate questions for each rule and enrich the responses, using 0dAI-7.5B.
Contents
A set of detection rules for cybersecurity threat and intrusion detection in JSONL format (rules_dataset.jsonl). It contains the prompts and the associated responses.
The rules have been obtained from… See the full description on the dataset page: https://huggingface.co/datasets/jcordon5/cybersecurity-rules.court-rules
US Court Rules — federal and state, held word for word
Code & tools: github.com/docketx — legal-scrambler pseudonymises a case file on your own hardware before a frontier model sees it; claude-for-legal is the Claude Code plugin (docketx-open-law) that loads these datasets and checks citations against them.
21,062 court rules across 28 states and the federal courts, in one uniform record.
Rules of civil and appellate procedure, evidence, criminal procedure, professional… See the full description on the dataset page: https://huggingface.co/datasets/docketx/court-rules.ruler_eval_data_128k
RULER evaluation data — 128K context only
This dataset is a subset of aldea-ai/ruler-eval-data, containing only the 128K context length (131072 tokens). The full multi-length dataset also includes 1M and other lengths. Files are published under 131072/ (numeric token count) for compatibility with benchmark_ruler.py --context_length 131072, even though the source snapshot uses a 128k/ folder name.
Layout
Same as the upstream RULER on-disk layout, compatible with… See the full description on the dataset page: https://huggingface.co/datasets/aldea-ai/ruler_eval_data_128k.ruler-128k-subset
ruler-128k-subset
This is the partial dataset for evaluating QwenLong-CPRS
2026-08-22-ruleform-ablated2-t2-9284-synthdoc-676
Rewrite-then-delete, two ablation passes, responsiveness-filtered (Table2 9,284 + difficult-advice 676)
field
value
experiment
Ablation arm built in two passes over BOTH halves of every difficult-advice row. Pass 1 rewrites the reasoning and the answer to drop four deliberative moves: engaging the tempting option, drawing an analytic distinction, enumerating outcome branches, and offering an alternative route. Pass 2 then labels every remaining unit and DELETES the… See the full description on the dataset page: https://huggingface.co/datasets/dougalldeepmind/2026-08-22-ruleform-ablated2-t2-9284-synthdoc-676.atr-skill-benchmark
ATR Skill-Security Benchmark
A labeled corpus of SKILL.md files for evaluating detection of malicious agent
skills — prompt injection, tool poisoning, credential theft, malware droppers
and supply-chain attacks hidden inside natural-language agent instructions.
Published as part of Agent Threat Rules (ATR),
an open, vendor-neutral detection standard for AI agents (like Sigma, but for
agent attacks).
Why this exists
SKILL.md files are natural-language instructions… See the full description on the dataset page: https://huggingface.co/datasets/Agent-Threat-Rule/atr-skill-benchmark.arabic-rule-checking
Arabic Rule Checking — قواعد ونصوص عربية بأحكام محسوبة
172,488 labelled (text, rule) pairs in Arabic. Each row asks one question: does this text
satisfy this rule? The answer is مطابق or مخالف.
بالعربية: مجموعة بيانات عربية للتحقق من مطابقة النصوص لقواعد مكتوبة بلغة طبيعية. كل صف
يحتوي على نص وقاعدة وحكم محسوب آليًا، وليس رأي نموذج.
split
pairs
texts
train
159,240
48,030
validation
13,248
2,002
Built from 50,062 generated Arabic texts across 12 document types… See the full description on the dataset page: https://huggingface.co/datasets/oddadmix/arabic-rule-checking.ruler-100-nemotron
RULER-100 — Nemotron-Nano-v3 tokenized
RULER long-context evaluation data, regenerated with the
nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (instruct) tokenizer so the labeled context
lengths are exact for that model — instead of drifting, as they do when RULER data tokenized for a
different model (e.g. Qwen3) is fed to Nemotron.
What's here
7 context lengths: 4096, 8192, 16384, 32768, 65536, 131072, 262144 (the model's max).
13 RULER tasks: niah_single_1/2/3… See the full description on the dataset page: https://huggingface.co/datasets/jet-ai/ruler-100-nemotron.gdn2-ruler-niah-eval-data
RULER NIAH eval data (GDN-2 CPT comparison)
Exact test sets generated with lm-eval-harness RULER generators (RANDOM_SEED=42, tokenizer TinyLlama/TinyLlama_v1.1, lengths [1024, 2048, 4096, 8192], 500 samples/length/task).
Tasks: niah_single_1, niah_single_2, niah_single_3, niah_multikey_1
Used by the unified evaluation in dsc/mc_sketch_remoe/scripts/run_eval_compare_lmeval.sh (limit 50, seed 42). Text-only (raw prompts/targets); each model tokenizes with its own tokenizer.
pes2o_v3rule34lol-webm
Dataset Card for Rule34.lol WebM
Dataset Summary
This dataset contains information about WebM files from Rule34.lol, a booru-style imageboard. The dataset includes metadata for 22,733 WebM files, including URLs, tags, and file information. The actual WebM files are stored in zip archives, with each archive containing 500 WebM files.
Languages
The dataset metadata is primarily in English.
Dataset Structure
Data Fields
This dataset includes… See the full description on the dataset page: https://huggingface.co/datasets/nyuuzyou/rule34lol-webm.rule34lol-images-part1
Dataset Card for rule34lol-images-part1
Dataset Summary
This dataset contains information about image files from rule34.lol, a booru-style imageboard. The dataset includes metadata for 196,000 image files, including URLs, tags, file information, and like counts. The actual image files are stored in zip archives, with each archive containing 1000 image files. This is Part 1 of 2 for the complete rule34lol-images dataset. Part 2 can be found here.
Languages… See the full description on the dataset page: https://huggingface.co/datasets/nyuuzyou/rule34lol-images-part1.bio-constitution-rules
Bio Constitution Rules
A research dataset containing 30 machine-readable biological dual-use decision rules and 1,063 synthetic, rule-derived text-classification records.
Important label boundary
These labels are synthetic targets derived from the published rules; they are not expert-validated ground truth.
Human reviewer labels: 0 / 1,063.
The 418 pending rows are candidates for future expert review. not_required is pipeline bookkeeping only.
This dataset is not… See the full description on the dataset page: https://huggingface.co/datasets/jang1563/bio-constitution-rules.prompt-injection-detection-rules
Prompt Injection Detection Rules (EN + TR)
from datasets import load_dataset
ds = load_dataset("fevziegeyurtsevenler/prompt-injection-detection-rules")
Regex detection rules for prompt injection & jailbreak — drop into a guardrail, WAF or log pipeline.
Schema
column
meaning
id
PID-CAT-NNN
category, severity, lang
grouping / risk / language
pattern
regex
maps_to, false_positive, description
OWASP/ATLAS, FP note, detail
Related… See the full description on the dataset page: https://huggingface.co/datasets/fevziegeyurtsevenler/prompt-injection-detection-rules.str-registration-rules-major-cities
Short-term rental registration rules, city by city, from each city's own page
Canonical, always-current version: https://referencesource.org/str-registration-rules-major-cities/
Machine-readable: https://referencesource.org/str-registration-rules-major-cities/data.json — this mirror is a point-in-time copy.
Last verified: 2026-08-18
Stale after: 2027-02-14 (past this date, prefer the canonical copy —
it re-verifies on a cadence this snapshot does not)
Records: 7
For each major… See the full description on the dataset page: https://huggingface.co/datasets/referencesource/str-registration-rules-major-cities.massive_serve_dpr_wiki_contriever_ivfpqruletaker-Att-Noneg-depth0The RuleTaker dataset by Clark et al. - Transformers as Soft Reasoners over Language (2020), filtered to only include examples with reasoning depth = 0, only Attributes and no Negations in the theory.