Invariants Are Necessary but Not Sufficient

Invariants stop structural corruption, not all meaningful regressions

Highlights

• The invariants guard is the first stop because fatal structural failures should fail fast.
• A pass on invariants only shows that basic shape, finiteness, and tying contracts held.
• Quality preservation requires later evidence from primary metrics and downstream guards.

What Invariants Actually Check

The public guard documentation is explicit that invariants are structural. They scan for non-finite tensors, tokenizer and embedding mismatches, broken weight tying, shape compatibility, and missing or drifting structural evidence such as LayerNorm or positional-encoding checks.

Early structural checking has a narrow job. Fatal invariant types such as non-finite tensors or tokenizer alignment failures block before evaluation. Other structural drift remains visible as warnings in default monitor mode unless strict mode or a blocking policy is configured.

Why A Pass Is Not A Quality Claim

Readers often blur structural sanity and acceptable behavior together. They are not the same thing.

An edit can preserve tensor shapes, avoid NaNs, and keep tied weights intact while still degrading the primary metric, tripping spectral limits, or violating activation stability. The guard contract primer makes that separation visible by listing invariants next to separate primary-metric, spectral, RMT, and variance gates.

Why Invariants Belong At The Front Of The Chain

The guard chain works because its stages do different jobs in sequence. Invariants clear the path for everything else. They answer: is this edit even well-formed enough to evaluate?

Putting invariants first is therefore not cosmetic. It is a resource and interpretation boundary. Fatal invariant failures should abort before later evidence is generated, because later evidence would only create noise around a structurally invalid artifact.

What The Later Guards Add

Once invariants pass, the rest of the pipeline starts to measure behavior rather than shape. Spectral monitoring checks baseline-relative weight drift. RMT checks activation edge-risk movement. The primary metric and paired evaluation windows check whether the edit actually preserved the intended task behavior.

A reviewer should therefore read an invariants pass as a necessary entry condition, not as the conclusion of the review.

Claim Map

The narrow claim is straightforward:

• invariants protect against structurally invalid edits
• invariants do not certify quality preservation
• later guards and primary metrics carry that burden

No single green check should carry the whole assurance story.

Limitations

• No new malformed-model case study is included.
• The claim is limited to the current public guard behavior: invariants are structural entry checks, while later guards and metrics carry the quality-preservation evidence.

Sources

• Guards
• Guard Contracts & Primer
• Assurance Case
• Reports Reference