Skip to content

Claim Validation Matrix

The claim validation matrix is the authoritative record of all empirical claims made by the Affective Field Theory (AFT) library, their current evidence status, and the references to benchmark artefacts that support or qualify each claim.

Each entry carries:

  • claim_text — the exact public-facing claim
  • requires_oracle_affect — true if the claim's primary positive-result benchmarks inject preset valence/arousal at encode time (oracle affect); false if the benchmark ran without oracle injection (naturalistic or appraisal-driven). Do not conflate oracle-affect results with end-to-end naturalistic results — they measure different things.
  • evidence_refs / benchmark_refs / protocol_refs / limitations_refs — file paths within the repository that serve as evidence (all paths validated in CI via tools/audit_claim_refs.py)
  • current_evidence — evidence classification (e.g. strong_intra_theory, strong_cross_task, not_yet_shown)
  • not_yet_shown — explicit scope limitations attached to each claim
  • allowed_public_wording — the precise wording permitted in communications
  • next_study — reference to any planned follow-up pre-registration

Download raw JSON{ .md-button }

Validation tool

uv run python tools/audit_claim_refs.py

Exits 0 if all referenced paths exist on disk; exits 1 with a summary of missing paths. Integrated into make meta-check and therefore into make check.