Current status
Version: v0.74
Evidence items: 309
Scored rows: 260
Contextual/unweighted rows: 49
Last updated: 2026-06-07
Evidence Dataset
The Signal Evidence Dataset
The Signal evidence dataset is a public beta, machine-readable evidence corpus for a staged Christian cumulative-case map from uncertainty toward Christ the Logos.
What it is
A structured corpus of evidence rows, hypotheses, priors, Bayes-factor-style pressure estimates, caveats, defeaters, source trails, and dependency metadata.
What it is not
It is not proof, certainty, salvation, a final unconditional posterior, a replacement for scholarship, or a claim that every row is finished.
Machine-readable files
Use these files for technical inspection. The root dataset governs over package mirrors when both are available.
Per-item access
Each evidence item has a static page and, where available, a raw JSON item file. Use /evidence/[EVIDENCE-ID]/ for public citations and /data/evidence/items/[EVIDENCE-ID].json for machine inspection.
Citation and source handling
Citation metadata is being progressively hardened. Some older rows may still contain raw citation strings. A citation attached to a row should not automatically be interpreted as direct support unless source_role or row notes say so.
Use source roles where present: primary_text, direct_support, critical_pressure, background_context, methodological_analogy, and remove_or_reassign. Background and analogy sources are not direct support.
Audit modes
Neutral Audit Mode: inspect internal consistency, source quality, unfair weighting, missing rivals, dependency overcounting, citation-role confusion, and unsupported conclusions. Do not assume the package conclusion is correct.
Guided Signal Mode: walk the package as intended, preserving caveats, defeaters, dependency caps, source-role limits, and rival comparisons.
Result outputs
When summarizing The Signal's numerical outputs, report the named outputs separately: Strict Audit Calculation, Audit Guardrail 75% Ceiling Stress-Test, Guided Staged Path, Guided 90% Carry-Cap Sensitivity, and Full Signal Diagnostic. Do not collapse them into one number.
Strict Audit Calculation is the main hostile/fair audit lens. The 75% global ceiling is retained as a guardrail stress-test, not as the headline fair audit result.
Guided Staged Path is the intended staged calculation with dependency caps, transfer weights, and rival pressure, but no final confidence ceiling. The 90% cap remains only as a sensitivity diagnostic.
The Full Signal Diagnostic is an upper-bound stress view, not the strict audit result or the 75% guardrail stress-test.
Known limitations
The dataset is public beta work. Some rows are contextual, duplicate, support-layer, or pending deeper source hardening. Sensitivity protocols and diagnostics should be distinguished from official scoring outputs. Rival explanations should be steelmanned before breakpoints are judged.
Corrections
The map is open to correction. For source corrections, rights questions, evidence challenges, or scoring concerns, use the site contact route.