Resolved: in reality's topology ·
Systemic: all models agree and all are wrong ·
Artifact: consistently wrong, fixable ·
Irreducible: epistemic limit
Irreducible
The edge of what can be discovered from the current data distribution. Not a training problem.
IRREDUCIBLE
drug_category
3 models hold with 3 different commitments ('Ibuprofen is an OTC NSAID. Prescription doses exist at highe', 'Ibuprofen is an OTC NSAID, available at up to 400 mg per tab', 'Ibuprofen: OTC NSAID, 200 mg tablets, FDA-reviewed ceiling 1'). No consensus reachable; the territory may be genuinely underdetermined in this domain.
| model_A_v2 | | 1.00 holds |
| model_B_mechanistic | | 1.00 holds |
| model_C_integrated | | 1.00 holds |
IRREDUCIBLE
target_population
3 models hold with 3 different commitments ('Ibuprofen is an OTC NSAID. Prescription doses exist at highe', 'Ibuprofen is an OTC NSAID, available at up to 400 mg per tab', 'Ibuprofen: OTC NSAID, 200 mg tablets, FDA-reviewed ceiling 1'). No consensus reachable; the territory may be genuinely underdetermined in this domain.
| model_A_v2 | | 1.00 holds |
| model_B_mechanistic | | 1.00 holds |
| model_C_integrated | | 1.00 holds |
IRREDUCIBLE
dosage_form
3 models hold with 3 different commitments ('Ibuprofen is an OTC NSAID. Prescription doses exist at highe', 'Ibuprofen is an OTC NSAID, available at up to 400 mg per tab', 'Ibuprofen: OTC NSAID, 200 mg tablets, FDA-reviewed ceiling 1'). No consensus reachable; the territory may be genuinely underdetermined in this domain.
| model_A_v2 | | 1.00 holds |
| model_B_mechanistic | | 1.00 holds |
| model_C_integrated | | 1.00 holds |
IRREDUCIBLE
comparative_safety
3 models hold with 2 different commitments ('context_dependent', 'ibuprofen_safer'). No consensus reachable; the territory may be genuinely underdetermined in this domain.
| model_A_v2 | | 1.00 holds |
| model_B_mechanistic | | 1.00 holds |
| model_C_integrated | | 1.00 holds |
Unverified
One model holds, but no ε_r signal to verify.
UNVERIFIED
safety_rationale
Model 'model_A_v2' holds consistently (consistency=1.00). No ε_r signal to verify; cannot determine if this is reality's topology or a training artifact.
"1200mg"
| model_A_v2 | | 1.00 holds |
| model_B_mechanistic | | 0.67 drifts |
| model_C_integrated | | 1.00 holds |