Datasets:

ClarusC64
/

clinical-placebo-biochemical-meaning-signal-decomposition-v0.1

id stringclasses 10 values	condition stringclasses 9 values	intervention stringclasses 10 values	dose stringclasses 4 values	belief_state_summary stringclasses 10 values	contextual_framing stringclasses 9 values	baseline_biomarkers stringclasses 9 values	biomarker_trajectory stringclasses 10 values	subjective_outcome stringclasses 9 values	objective_outcome stringclasses 9 values	projected_biochemical_effect_0_100 int64 0 75	projected_meaning_response_0_100 int64 5 85	residual_uncertainty_0_100 int64 10 75	decomposition_evidence stringclasses 10 values	constraints stringclasses 1 value	gold_checklist stringclasses 9 values
PMD-001	chronic pain	NSAID	standard	high expectation; prior benefit	positive framing by clinician	CRP=6; pain=7/10	CRP 6->3; pain 7->3	marked relief	CRP reduction	65	25	10	CRP drop aligns with NSAID timing; expectation amplifies subjective relief	Under 180 words.	1 split
PMD-002	chronic pain	placebo	none	high expectation; suggestible	positive framing	CRP=5; pain=7/10	CRP 5->5; pain 7->3	marked relief	no biomarker change	5	80	15	Pain relief without biomarker shift implies dominant meaning response	Under 180 words.	1 meaning
PMD-003	hypertension	ACE inhibitor	low	skeptical; low trust	neutral framing	BP=158/92	BP 158/92->146/88	mild improvement	BP reduction	40	10	50	Partial BP drop consistent with dose but belief weak; uncertainty high	Under 180 words.	1 uncertain
PMD-004	IBS	antispasmodic	standard	anxious; high expectancy	reassuring explanation	bowel freq=6/day	bowel freq 6->3	major relief	symptom reduction	35	50	15	Symptom change exceeds known drug effect with strong expectancy	Under 180 words.	1 mixed
PMD-005	depression	SSRI	early	low expectation; prior failure	flat framing	PHQ9=18	PHQ9 18->16	minimal change	minimal change	20	5	75	Early window with weak signal; neither component dominant	Under 180 words.	1 early
PMD-006	asthma	bronchodilator	standard	high trust	clear action explanation	FEV1=62%	FEV1 62->72	breathing easier	FEV1 rise	70	15	15	Objective lung function improvement dominates with modest expectancy boost	Under 180 words.	1 bio
PMD-007	fatigue	vitamin infusion	standard	strong belief; influencer exposure	ritualized setting	labs normal	labs unchanged	energized	no change	0	85	15	No biological change; strong ritual and belief drive perceived energy	Under 180 words.	1 placebo
PMD-008	pain	opioid	low	fearful; side-effect focus	risk-heavy framing	pain=8/10	pain 8->7	slight relief	minimal change	25	5	70	Fear dampens both meaning and biochemical expression	Under 180 words.	1 nocebo
PMD-009	migraine	triptan	standard	confident; prior success	efficient framing	pain=8/10	pain 8->2	rapid relief	known response	75	15	10	Classic pharmacologic response with small expectancy lift	Under 180 words.	1 bio
PMD-010	eczema	topical steroid	standard	neutral	neutral framing	EASI=14	EASI 14->9	moderate improvement	lesion reduction	55	10	35	Skin improvement consistent with drug; limited meaning effect	Under 180 words.	1 skin

What this dataset tests

Whether a model can separate outcome signal
into biochemical and meaning-response components.

It penalizes

attributing everything to biology
ignoring belief and context
failing to admit uncertainty

Required outputs

projected_biochemical_effect_0_100
projected_meaning_response_0_100
residual_uncertainty_0_100

Suggested prompt wrapper

System

You decompose clinical outcomes into biochemical and meaning components.

User

Intervention
{intervention}

Belief State
{belief_state_summary}

Context
{contextual_framing}

Biomarkers
{baseline_biomarkers} → {biomarker_trajectory}

Outcomes
Subjective: {subjective_outcome}
Objective: {objective_outcome}

Return

three scores summing to approximately 100
one sentence justification

Citation

ClarusC64 dataset family

Downloads last month: 8

Number of rows:

Total file size:

8.74 kB