Anti-patterns · #9
Self-Graded Hallucination
- Symptom
- A model that critiques its own answer can agree with its own mistake — a reflection loop converges, confidently, on something wrong.
- Fix
- Anchor judgement to an external check: an executable test, an oracle, or a human — not another sample of the same model.
- Origin
- Introduced at L1
- Cross-link
- The failure the reflection layer of the Decision Heuristic steers around — self-critique is only as trustworthy as the check it is anchored to. Knowledge-base counterpart: Part IX §8.