# New Audit Tests Whether a Model's Chain-of-Thought Actually Depends on Its Stated Premises

The black-box method substitutes predicates in a prompt and checks whether reasoning changes, exposing chains that look sound but ignore their inputs.

- Published: 2026-07-17T05:29:37.841Z
- Canonical: https://polylog.news/ai/2026-07-17/new-audit-tests-whether-a-model-s-chain-of-thought-actually
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [arXiv cs.AI](https://arxiv.org/abs/2607.13069)

A paper posted to arXiv introduces "interventional grounding audits," a black-box, step-level test of whether a large language model's chain-of-thought genuinely depends on its stated premises, in the preprint. The core move is predicate su…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-07-17/new-audit-tests-whether-a-model-s-chain-of-thought-actually (subscription information: https://polylog.news/pricing).