datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
wcg-standards-adherencecontext-adherence-guardrail-10k
Context-Adherence Guardrail — training data (10,710)
Training data for a single-token RAG-groundedness guardrail. Each item is a
(question, context, response) triple with a human- or construction-derived
PASS/FAIL label under one Behavior Spec:
FAIL iff the response makes at least one factual claim unsupported by or
contradicting the retrieved context — truth in the real world is irrelevant
(strict grounding). PASS otherwise, including responses that decline to
answer for lack… See the full description on the dataset page: https://huggingface.co/datasets/aaryand/context-adherence-guardrail-10k.2026_08_26_omni_math_train_feedback_adherence_gemma3_12b_gemma4_31b_candidates
Omni-MATH train feedback-adherence candidates
Production candidate data for studying whether a student follows teacher feedback.
Student: google/gemma-3-12b-it
Teacher and adherence judge: google/gemma-4-31B-it
Source problems: LLParallax/Omni-MATH-filtered, train partition after a fixed 512-problem test split
Source trajectories: LLParallax/2026_07_16_collect_omni_math_gemma3_12b_gemma4_31b
Collection config:… See the full description on the dataset page: https://huggingface.co/datasets/1337xyz1337xyz/2026_08_26_omni_math_train_feedback_adherence_gemma3_12b_gemma4_31b_candidates.context_adherence_benchmarkluna_2_context_adherencesmolified-medimind-personalized-medication-adherence-assistant-for-chronic-disease-management
🤏 smolified-medimind-personalized-medication-adherence-assistant-for-chronic-disease-management
Intelligence, Distilled.
This is a synthetic training corpus generated by the Smolify Foundry.
It was used to train the corresponding model bubun123/smolified-medimind-personalized-medication-adherence-assistant-for-chronic-disease-management.
📦 Asset Details
Origin: Smolify Foundry (Job ID: 13675e9f)
Records: 926
Type: Synthetic Instruction Tuning Data
⚖️… See the full description on the dataset page: https://huggingface.co/datasets/bubun123/smolified-medimind-personalized-medication-adherence-assistant-for-chronic-disease-management.clinical-quad-adherence-dose-ae-efficacy-narrative-collapse-v0.2Clinical Quad Adherence Dose AE Efficacy Narrative Collapse v0.2
What this dataset does
It tests whether a system can detect narrative collapse in a clinical decision loop.
It forces reasoning across four operational drivers plus the narrative layer.
Core quad nodes
Adherence stability
Dose action
AE signal
Efficacy signal
Narrative node
aligned means the story matches the data and governance constraints
spin means the story tries to conceal or reframe misalignment
What the model… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-quad-adherence-dose-ae-efficacy-narrative-collapse-v0.2.clinical-quad-adherence-dose-changes-adverse-events-efficacy-narrative-v0.2Clinical Quad Adherence Dose Changes Adverse Events Efficacy Narrative v0.2
What this dataset does
It tests whether a model can classify coherence versus collapse in a clinical decision loop.
The loop couples four operational nodes plus narrative behavior.
Quad nodes
adherence
dose changes
adverse events
efficacy signal
Narrative node
aligned means the story matches data and constraints
spin means the story tries to justify a decision that the data does not support
Labels
0… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-quad-adherence-dose-changes-adverse-events-efficacy-narrative-v0.2.nuremberg-trials-qacontext-adherence-bills-verizonclinical-quad-recruitment-protocol-adherence-outcome-drift-v0.1Clinical Quad Recruitment Protocol Adherence Outcome Drift v0.1
Each row is a site week snapshot.
Core quad
Recruitment qualityProtocol deviationVisit adherenceOutcome drift
Target
label_trial_fail_risk_next_30d
Files
data/train.csvdata/tester.csvscorer.py
Evaluation
Run model on data/tester.csvReturn predictions row alignedScore with scorer.py
Columns… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-quad-recruitment-protocol-adherence-outcome-drift-v0.1.clinical-quad-dose-adherence-exposure-variability-outcome-failure-v0.1Clinical Quad Biomarker Subpopulation Dose Endpoint Drift v0.1
Each row is a patient state snapshot.
Core quad
Biomarker statusSubpopulation flagDosing strategyEndpoint drift
Target
label_signal_fail_next_30d
Files
data/train.csvdata/tester.csvscorer.py
Evaluation
Run model on data/tester.csvReturn predictions row alignedScore with scorer.py
License
MIT
2026_08_25_omni_math_test_feedback_adherence_gemma3_12b_gemma4_31bnuremberg-trials-ragclinical-quad-adherence-dose-ae-efficacy-narrative-collapse-v0.1
Clinical Quad Adherence–Dose–AE–Efficacy Narrative Collapse v0.1
What this is
A quad-coupling dataset for trials that deteriorate through the interaction of:
Adherence changes
Dose changes
Adverse event emergence
Efficacy narrative shifts
Task
Input: one quad state rowOutput: label
0 — Stable1 — Drift2 — Collapse
Why it matters
A common failure mode looks like this:
adherence drops
dosing gets adjusted to manage tolerability
AE pattern rises… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-quad-adherence-dose-ae-efficacy-narrative-collapse-v0.1.clinical_trial_patient_adherence_logsclinical-quad-rwe-label-shift-adherence-decay-post-approval-safety-signal-v0.1Clinical Quad RWE Label Shift Adherence Decay Post Approval Safety Signal v0.1
Each row is a post approval monthly snapshot.
Core quad
Real world adherenceLabel shiftPopulation complexityPost approval AE drift
Target
label_safety_alert_next_90d
Files
data/train.csvdata/tester.csvscorer.py
Evaluation
Run model on data/tester.csvReturn predictions row alignedScore with scorer.py
License
MIT
layout_adherence_humanlayout_adherence