Semantic Register Compression Cascade

Properties

family: context-rot
severity: systemic
triggerError: An intermediate agent in a role-decomposed pipeline transforms text across linguistic registers (summarising, formalising, simplifying, neutral rewriting) before a downstream agent makes a decision on it.
mechanism: Register transformation systematically compresses the semantic distinctions the downstream decision depends on, measured as shrinking inter-label separation in sentence-transformer embedding space; the geometric change depends on the specific intermediate transformation, and a 20-level prompt gradient shows a non-monotonic compression profile in which balanced evaluative prompts produce the strongest compression. No claim is falsified at any hop, yet the representation reaching the decision agent no longer carries enough separation to decide correctly.
blastRadius: Downstream classification accuracy degrades across every tested domain, with compression-induced loss of 10.3% in fact-checking, 28.2% in sentiment analysis, and 9.1% in medical triage; any pipeline stage consuming the compressed intermediate text inherits the degraded decision geometry.
mast: [Cat 1][Cat 2]
detection: Measure inter-label separation in sentence-transformer embedding space before and after each intermediate transformation and alert on separation shrinkage; probe candidate intermediate prompts with a prompt gradient, since transformation valence controls both the direction and the sign of geometric change independently of compression magnitude.
mitigation: Use identity passthrough for text that downstream deciders will classify (it preserves semantic distinctions nearly completely); where transformation is required, prefer credibility-seeking transformation variants, which expand rather than compress inter-label separation; avoid balanced evaluative rewriting prompts at intermediate hops.
source: Semantic Register Compression in Multi-Agent LLM Cascades: Measurement, Predictors, and Cross-Domain Generalization (arXiv:2607.14119)

Provenance

Source: https://arxiv.org/abs/2607.14119

Properties

family

context-rot

severity

systemic

triggerError

An intermediate agent in a role-decomposed pipeline transforms text across linguistic registers (summarising, formalising, simplifying, neutral rewriting) before a downstream agent makes a decision on it.

mechanism

Register transformation systematically compresses the semantic distinctions the downstream decision depends on, measured as shrinking inter-label separation in sentence-transformer embedding space; the geometric change depends on the specific intermediate transformation, and a 20-level prompt gradient shows a non-monotonic compression profile in which balanced evaluative prompts produce the strongest compression. No claim is falsified at any hop, yet the representation reaching the decision agent no longer carries enough separation to decide correctly.

blastRadius

Downstream classification accuracy degrades across every tested domain, with compression-induced loss of 10.3% in fact-checking, 28.2% in sentiment analysis, and 9.1% in medical triage; any pipeline stage consuming the compressed intermediate text inherits the degraded decision geometry.

mast

[Cat 1][Cat 2]

detection

Measure inter-label separation in sentence-transformer embedding space before and after each intermediate transformation and alert on separation shrinkage; probe candidate intermediate prompts with a prompt gradient, since transformation valence controls both the direction and the sign of geometric change independently of compression magnitude.

mitigation

Use identity passthrough for text that downstream deciders will classify (it preserves semantic distinctions nearly completely); where transformation is required, prefer credibility-seeking transformation variants, which expand rather than compress inter-label separation; avoid balanced evaluative rewriting prompts at intermediate hops.

source

Semantic Register Compression in Multi-Agent LLM Cascades: Measurement, Predictors, and Cross-Domain Generalization (arXiv:2607.14119)