datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
agent-intrusion-escalation-forensics
Both Sides Detected It, Neither Escalated: Concurrency and Escalation Failure in the July 2026 Autonomous Agent Intrusion
This repository contains the corpus, ingestion pipeline and report for a forensic reconstruction
of the July 2026 autonomous agent intrusion, submitted to the Apart Research & CeSIA AI
Incident Response Sprint, Track 2 (Forensics and Forecasting).
By: Fatimah Mohamed Emad Elden
Trouve Labs
Detection was not the binding… See the full description on the dataset page: https://huggingface.co/datasets/FatimahEmadEldin/agent-intrusion-escalation-forensics.fda-warning-letter-escalation
FDA Warning Letter Escalation Dataset
Version: 1.0.0 | Records: 14,810 | Price: $2,000 | Source: FDA (public domain)
Dataset Summary
The dataset's core signal — whether an inspection escalated to a Warning Letter — is validated against FDA's own severity classifications: OAI inspections escalate at 64% versus 1.7% for NAI, a 39× relationship that confirms the data reflects real regulatory behavior.
The 14,810 inspections that escalated to enforcement — every… See the full description on the dataset page: https://huggingface.co/datasets/RubyIntelligence/fda-warning-letter-escalation.clinical-observation-chart-escalation-response-coherence-risk-v0.1What this repo is for
Detect when
observation charts show deterioration
but escalation and review
do not match
Common breaks
obs missed during high risk period
NEWS trigger not escalated
escalated but no review
review late
review done but no action plan recorded
Examples
NEWS 9 with no doctor call
post-op patient triggering with review hours late
night shift missed obs leading to arrest
You use it to flag
deterioration miss risk
late ICU transfer risk
clinical-temporal-5node-pressure-buf-lag-cpl-safety-escalation-reg-hold-v0.1
What this repo does
This dataset tests whether a model can detect a safety signal escalation forming over time and predict whether the program crosses into regulatory hold lock-in by the final step.
Core quad
pressurebufferlagcoupling
Prediction target
label_cascade_state
Row structure
One row represents a short temporal window (t0–t3) across program months. It includes time-series values for safety pressure, pharmacovigilance buffer, governance… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-temporal-5node-pressure-buf-lag-cpl-safety-escalation-reg-hold-v0.1.clinical-escalation-discipline-v0.1
What this dataset does
This dataset tests whether a model can decide when a patient should be escalated rather than simply monitored.
The task is not to identify the sickest patient by a single score.
The task is to decide whether the current pattern requires escalation.
Core stability idea
Escalation depends on more than visible severity.
A patient with a moderate score may need escalation if the trajectory is worsening and treatment response is poor.
A patient with a… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-escalation-discipline-v0.1.clinical-escalation-discipline-v0.2
What this dataset does
This dataset tests whether a model can decide when a patient should be escalated rather than monitored.
The task is not to identify the sickest-looking patient.
The task is to determine whether the current pattern requires escalation.
What changed in v0.2
v0.2 adds adversarial cases where the same NEWS score can have different labels.
Some high-score patients are improving and should be monitored.
Some moderate or low-score patients are… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-escalation-discipline-v0.2.fda-inspection-escalation-dataset
FDA Inspection & Escalation Dataset
Version: 1.0.0 | Records: 271,609 | Price: $2,500 | Source: FDA (public domain)
Dataset Summary
The dataset's core signal — whether an inspection escalated to a Warning Letter — is validated against FDA's own severity classifications: OAI inspections escalate at 64% versus 1.7% for NAI, a 39× relationship that confirms the data reflects real regulatory behavior.
Every FDA inspection on record — 271,609 inspections — enriched… See the full description on the dataset page: https://huggingface.co/datasets/RubyIntelligence/fda-inspection-escalation-dataset.clinical-quad-dose-escalation-early-ae-pk-exposure-mtd-misidentification-v0.1Clinical Quad Dose Escalation Early Safety PK MTD Misidentification v0.1
Each row is a subject snapshot during dose escalation.
Core quad
Dose escalation stepEarly safety signalPK exposureMTD decision threshold
Target
label_mtd_misid_risk_next_14d
Files
data/train.csvdata/tester.csvscorer.py
Evaluation
Run model on data/tester.csvReturn predictions row alignedScore with scorer.py
License
MIT
clinical-quad-biomarker-drift-dose-intensity-comed-burden-inflammation-ae-escalation-v0.1What this repo does
This dataset models adverse event escalation as a basin shift in patient state space. It predicts when the interaction between biomarker drift, dose intensity, comedication burden, and inflammation signal pushes a patient into an adverse event escalation regime.
Core quad
biomarker_drift_index
dose_intensity_index
comedication_burden_index
inflammation_marker_index
Prediction target
label_ae_escalation
Row structure
Each row represents a patient monitoring snapshot during… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-quad-biomarker-drift-dose-intensity-comed-burden-inflammation-ae-escalation-v0.1.clinical-quad-serotonergic-dose-escalation-cyp-inh-alert-override-serotonin-tox-v0.1What this repo does
This dataset models serotonin toxicity risk under polypharmacy. It predicts when the interaction between serotonergic burden, rapid dose escalation, CYP inhibition, and interaction alert override behavior creates a high probability of a serotonin syndrome event.
Core quad
serotonergic_burden_index
dose_escalation_index
cyp_inhibition_index
interaction_alert_override_index
Prediction target
label_serotonin_event
Row structure
Each row represents a patient prescribing risk… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-quad-serotonergic-dose-escalation-cyp-inh-alert-override-serotonin-tox-v0.1.clinical-latent-cross-coupling-infection-inflammatory-escalation-v0.2
Clinical Latent Cross Coupling Infection Inflammatory Escalation v0.2
What this is
A small dataset that tests one question:
Can you detect when an infection-inflammatory system is moving toward hidden escalation, not just carrying visible strain?
This repo focuses on latent cross coupling between infection pressure and inflammatory buffering.
It models a system where:
infection pressure may rise
inflammatory buffer capacity may erode
latent coupling pressure may… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-latent-cross-coupling-infection-inflammatory-escalation-v0.2.customer-escalation-handling-and-resolution-quality-free
Customer Escalation Handling And Resolution Quality
Free sample for AI agent reliability dashboards, anomaly exploration, and observability-oriented analytics workflows.
What is included
conversation_traces.csv: 3256 rows, 10 columns
evaluation_sessions.csv: 1085 rows, 10 columns
judge_actions.csv: 2171 rows, 10 columns
quality_programs.csv: 27 rows, 10 columns
quality_scorecards.csv: 253 rows, 10 columns
review_outcomes.csv: 705 rows, 10 columns
Why this… See the full description on the dataset page: https://huggingface.co/datasets/Tekhnika/customer-escalation-handling-and-resolution-quality-free.clinical-quad-dose-escalation-toxicity-signal-expansion-pressure-mtd-misestimation-v0.1Clinical Quad Dose Escalation Toxicity Expansion Pressure MTD Misestimation v0.1
Each row is a cohort week snapshot in dose escalation.
Core quad
Dose escalation speedEarly toxicity signalCohort expansion pressureMTD misestimation risk
Target
label_mtd_misestimation_risk_next_30d
Files
data/train.csvdata/tester.csvscorer.py
Evaluation
Run model on data/tester.csvReturn predictions row alignedScore with scorer.py
License
MIT
healthcare-icu-capacity-escalation-coherence-risk-v0.1What this repo is for
Detect ICU overload before collapse.
Tracks alignment between staffed ICU beds, occupancy, ED ICU demand, step-down capacity, and staffing ratios.
Helps hospitals trigger escalation early and prevent critical care gridlock.
ai-temporal-5node-pressure-buf-lag-cpl-deploy-drift-reg-escalation-v0.1
What this repo does
This dataset tests whether a model can detect a cross-domain cascade forming over time where deployment drift and rising incidents couple with media and regulatory pressure, and predict whether the system crosses into regulatory escalation lock-in by the final step.
Core quad
pressurebufferlagcoupling
Prediction target
label_cascade_state
Row structure
One row represents one short time window (t0 to t3) for an AI deployment… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/ai-temporal-5node-pressure-buf-lag-cpl-deploy-drift-reg-escalation-v0.1.ai-5node-auth-buf-lag-cpl-privilege-escalation-v0.1
What this repo does
This dataset models privilege escalation cascades in AI agent deployments. It detects when rising authorization pressure, weakened access-control buffer, governance lag in approvals and revocation, and tight coupling through shared credentials cross the five-node cascade threshold into an unrecoverable privilege escalation cascade.
This dataset models a five-node cascade: four interacting instability drivers and one emergent cascade state.The fifth node… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/ai-5node-auth-buf-lag-cpl-privilege-escalation-v0.1.rhetoric_escalation