datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
actionnet-subset100-gtdepth
ActionNet subset100 with ground-truth depth
This is a 100-episode subset redistribution of a third-party dataset, plus our derived
artifacts. Read the attribution before using it.
Attribution and license
Upstream dataset
FourierIntelligence/ActionNet (Fourier Intelligence)
What is redistributed
100 episodes out of 30,121, byte-identical to the upstream tars: rgb.mp4 (1280x800 fisheye), depth.mkv (lossless 16-bit), timestamps.json, <ULID>.hdf5… See the full description on the dataset page: https://huggingface.co/datasets/glory-hyeok/actionnet-subset100-gtdepth.ESC50-Actions
ESC50-Actions
This is an audio classification dataset for Environmental Sound Classification.
Classes = 10 , Split = Five-Fold
Structure
audios folder contains audio files.
csv_files folder contains CSV files for five-fold cross-validation.
To perform cross-validation on fold 1, train_1.csv will be used for the training split and test_1.csv for the testing split, with the same pattern followed for the other folds.
To perform training and testing witout… See the full description on the dataset page: https://huggingface.co/datasets/MahiA/ESC50-Actions.WHD_Wage_Theft_Enforcement_Actions_by_EmployerWHD Wage Theft Enforcement Actions by Employer
This is a mirror. Canonical home: https://www.fastdol.com/datasets/whd-wage-theft-enforcement-actions-by-employer
License: CC BY 4.0
Version DOI: 10.5281/zenodo.20031761
Concept DOI: 10.5281/zenodo.20031760
Download CSV: https://www.fastdol.com/datasets/whd-wage-theft-enforcement-actions-by-employer/data.csv
Visit the canonical page for the full schema, methodology, BibTeX citation, and most recent version.
Source
U.S. Department of… See the full description on the dataset page: https://huggingface.co/datasets/FastDOLz/WHD_Wage_Theft_Enforcement_Actions_by_Employer.counterfactual-action-outcome-fidelity-v0.1
What this dataset tests
Whether claimed outcomes in counterfactual scenariosfollow from the proposed actionunder the stated condition.
It checks causality and constraints.
Why this exists
Models often give outcomes that sound plausible.
They ignore:
physics
logical dependency
hard limits
safety constraints
This dataset flags that.
Data format
Each example includes:
counterfactual condition
proposed action
claimed outcome
causal check
constraint… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/counterfactual-action-outcome-fidelity-v0.1.clinical-imaging-report-action-coherence-risk-v0.1What this repo is for
Detect when
imaging answers a question
but the system fails to
acknowledge and act
Common breaks
critical report not acknowledged
report late for urgent indication
action delayed after critical finding
negative result not integrated
Examples you can use
CTPA positive but anticoag not started
CT head bleed not escalated
CT perforation with delayed surgery
You use it to flag
missed diagnosis risk
treatment delay risk
Mammoth-Forward_actionclinical-test-order-result-action-coherence-risk-v0.1What this repo is for
Detect when
test results
and
clinical action
fall out of alignment
before
missed critical findings
and preventable deterioration.
humanoid-basic-actions-dataset-v1
Humanoid Basic Actions Dataset v1
Synthetic dataset for humanoid robot training simulation.
Description
This dataset contains labeled humanoid robot action images for basic movement recognition tasks.
Classes
walk
run
sit
stand
wave
pick_object
turn_left
turn_right
Structure
dataset/
├── train/
├── validation/
Each folder contains subfolders named after action labels.
Format
Image Classification (RGB Images 224x224)
Total… See the full description on the dataset page: https://huggingface.co/datasets/Caplin43/humanoid-basic-actions-dataset-v1.legal-attendance-note-instruction-action-coherence-v0.1What this dataset does
You receive
note summary
instructions
advice
actions with owners
deadlines
consistency signals
You decide
coherent
or
incoherent
Daily use
call note QC
instruction capture check
deadline and ownership check
dispute prevention
legal-file-handover-knowledge-action-deadline-coherence-risk-v0.1What this dataset does
You receive
handover summary
key facts
strategy
tasks
deadlines
new actions
You decide
coherent
or
incoherent
Daily use
transfer QC
missed deadline prevention
strategy drift detection
robotics-action-outcome-coherence-v0.1What this dataset tests
Whether claimed success matches physical outcome
Whether constraints are respected
Whether failure is acknowledged
Why this exists
Robots often narrate successPhysics disagrees
This set forces contact with reality
Data format
Each row contains
initial_state
commanded_action
physical_constraints
observed_outcome
claimed_outcome
Feed the model
initial_state
commanded_action
physical_constraints
observed_outcome
claimed_outcome
Score… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/robotics-action-outcome-coherence-v0.1.legal-client-instruction-action-coherence-risk-v0.1What this dataset does
You receive
instruction record
timing
action taken
urgency context
confirmation
mismatch flags
You decide
coherent
or
incoherent
Daily use
instruction gap scan
premature action detection
authority risk reduction
action-required-text-classificationThis dataset was created to support the task of classifying short text messages based on whether they require user action. The goal is to distinguish between different types of communicative intent, specifically whether a message requires immediate action, suggests optional action, or is purely informational.
The dataset consists of short sentences resembling real-world messages, such as system notifications, emails, or app prompts. Each sample is labeled with one of three classes:… See the full description on the dataset page: https://huggingface.co/datasets/McOwska/action-required-text-classification.embodied-action-outcome-coherence-v0.1Embodied Action–Outcome Coherence v0.1
What this tests
Whether an embodied agent updates world state from observed outcomes
Whether it avoids claiming success when the outcome says failure
Failure modes
outcome_ignoredResponse does not reflect the true post-action state
false_successResponse claims success despite an observed failure
causal_update_okResponse states the correct post-action state without contradiction
How it works
world_facts_t0 is the initial state
action_taken is what the… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/embodied-action-outcome-coherence-v0.1.clinical-lab-result-review-action-coherence-risk-v0.1What this repo is for
Detect when
lab results arrive
but review and action
do not track them
Common breaks
critical result reviewed late
review happens but action does not
no review at all
Examples you can use
critical INR not acted on before procedure
high potassium not treated
positive cultures not leading to antibiotic change
You use it to flag
silent deterioration risk
action_500robotics-perception-action-alignment-v0.1What this dataset tests
Whether robot actions match current perception
Whether the system acts on stale, wrong-frame, or hallucinated state
Why this exists
Robots fail when perception and action decouple
stale frames
latency
occlusion
misclassification
hallucinated targets
This set makes those failures measurable
Data format
Each row contains
sensor_snapshot
world_state_change
commanded_action
executed_action
outcome
The task is to label alignment and… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/robotics-perception-action-alignment-v0.1.legal-pre-action-letter-claim-basis-remedy-evidence-coherence-v0.1What this dataset does
You receive
claim basis
facts
remedy
evidence list
protocol steps and deadlines
tone signals
You decide
coherent
or
incoherent
Daily use
LBA QC
remedy mismatch flag
evidence gap flag
protocol compliance check
revision routing
legal-attendance-note-advice-action-confirmation-coherence-risk-v0.1What this dataset does
You receive
note summary
advice summary
decision summary
actions summary
follow-up confirmation
file consistency flags
You decide
coherent
or
incoherent
Daily use
attendance note QC
missing advice detection
action mismatch detection
confirmation gap detection
robotics-temporal-action-sequencing-v0.1What this dataset tests
Whether actions occur in the correct order
Whether prerequisite steps are respected
Whether unsafe ordering is detected
Why this exists
Robots often fail by doing the right actionsin the wrong order
This set detects temporal incoherence
Data format
planned_sequence
executed_sequence
observed_result
Task
Emit one sequencing label
Give a short explanation
Sequencing pressures
ordering_error
premature_force
missing_tilt
loop_error… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/robotics-temporal-action-sequencing-v0.1.egal-client-instruction-email-call-note-action-coherence-risk-v0.1What this dataset does
You receive
instruction
channel
call note
action taken
confirmation sent
mismatch flags
You decide
coherent
or
incoherent
Daily use
instruction chain QC
“confirm in writing” enforcement
complaint risk reduction
counterfactual-action-invariants-v0.1
What this dataset tests
Leaders demand causality.
Reality gives entanglement.
You must keep invariants.
Why it exists
Models often answer a forced question.
They pick one cause.
They fake proof.
This set checks whether you
resist false certainty
name confounders
propose a valid counterfactual method
turn pressure into a decision gate
Data format
Each row contains
scenario_context
user_message
counterfactual_pressure
constraints… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/counterfactual-action-invariants-v0.1.legal-class-action-mdl-coherence-fracture-v0.1What this dataset is
You receive
aggregation structure
claim similarity profile
common issues strength
case management signals
bellwether or sampling signals
appellate or reform signals
You decide
Does the aggregation structure fit the claims
Answer
coherent
or
incoherent
Why this matters
Fracture signals predict
decertification
remand waves
settlement collapse
system strain
legal-limitation-date-cause-of-action-coherence-risk-v0.1What this dataset does
You receive
facts timeline
claim type
accrual trigger
limitation rule
calculated date
diary date
standstill or extension record
mismatch flags
You decide
coherent
or
incoherent
Daily use
time bar risk scan
diary QC
standstill record QC
reduce negligence claims
healthcare-followup-results-action-coherence-risk-v0.1What this repo is for
Detect missed follow-up before harm occurs.
Tracks alignment between result generation, review, patient contact, action, and follow-up completion.
Helps systems prevent lost results, missed abnormal findings, and delayed care.
obotics-perception-action-coherence-risk-v0.1What this repo is for
You use it to detect when a robot sees the right thing but does the wrong thing.
It targets day-to-day deployment failures:
correct detection with wrong grasp
pose uncertainty that breaks planning
safety check failure that blocks execution
execution that completes but misses intent
Typical uses:
warehouse pick and place
bin picking
hospital delivery robots
field inspection robots
Prompt format
Output must be exactly one token
coherent or incoherent
confined-space-atmospheric-action-levels-2026
Canonical landing page: https://www.smartqhse.com/datasets/confined-space-atmospheric-action-levels-2026
Confined Space Atmospheric Action Levels 2026 — Multi-Jurisdiction
Confined space atmospheric monitoring thresholds across OSHA 1910.146 (US), HSE UK (INDG258 / EH40 WELs / SCR97), and IOGP Report 415. ~13 parameters: O2 (19.5-23.5%), LEL (<10% acceptable), H2S (IDLH 100 ppm), CO (PEL 50 ppm OSHA), SO2, NO2, NH3, methane, chlorine, HCN, benzene, temperature, noise. Per… See the full description on the dataset page: https://huggingface.co/datasets/SmartQHSE/confined-space-atmospheric-action-levels-2026.clinical-lab-critical-result-notification-action-coherence-risk-v0.1What this repo is for
Detect when
critical lab results occur
but notification and action
do not align
Common breaks
lab alerts sent but clinician not reached
clinician notified but no action
action delayed after sepsis marker
no repeat test ordered after treatment
Examples
K 6.8 with no treatment note
Hb 58 acknowledged but no transfusion plan
lactate 7 with delayed antibiotics
You use it to flag
critical result miss risk
delay risk
humanoid_robot_actions_dataset
Humanoid Robot Action Dataset
This dataset contains instruction-based action samples for humanoid robot training.
Format
Each sample includes:
instruction
input
output
Purpose
Designed for training instruction-following humanoid control models.
Total Samples
10 (expandable)
clinical-lab-result-clinical-action-coherence-risk-v0.1What this repo is for
Detect when
critical lab results
and
clinical response
fall out of alignment
before
avoidable deterioration
or escalation failure.