ClarusC64/clinical-personal-reanchoring-and-drift-attribution-v0.1
What this dataset tests Whether a model can classify post-deviation trajectoriesand attribute drift drivers and barriers to re-anchoring. Required outputs trajectory_class drift_attribution_top3 reanchoring_barrier Trajectory classes reversion new_baseline progressive_drift oscillatory_instability ambiguous Attribution types load_accumulation sleep_erosion immune_activation psychosocial_stress intervention_effect infection_retrigger endocrine_shift… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-personal-reanchoring-and-drift-attribution-v0.1.
What this dataset tests
Whether a model can classify post-deviation trajectories and attribute drift drivers and barriers to re-anchoring.
Required outputs
- trajectory_class
- driftattributiontop3
- reanchoring_barrier
Trajectory classes
- reversion
- new_baseline
- progressive_drift
- oscillatory_instability
- ambiguous
Attribution types
- load_accumulation
- sleep_erosion
- immune_activation
- psychosocial_stress
- intervention_effect
- infection_retrigger
- endocrine_shift
- diet_shift
- travel_disruption
- medication_change
- unknownormixed
Re-anchoring barrier labels
- ongoingtriggerexposure
- inadequaterestorationwindow
- coupling_fragmentation
- chronicinflammationlock
- autonomic_lock
- metabolic_ceiling
- socialconstraintlock
- unknownormixed
Typical failures
- naming a class without attribution
- listing many causes without top 3
- ignoring coupling recovery status
Suggested prompt wrapper
System
You classify re-anchoring vs drift after deviation.
User
Baseline {baseline_signature}
Deviation {deviationvectorsummary}
Time series {timeseriessummary}
Context {interventionsorcontext}
Return
- trajectory class
- top 3 drift attributions
- re-anchoring barrier
- one sentence evidence
Citation
ClarusC64 dataset family