Metraipolis › archive › dc_7800596dca71

The Hallucination Snowball: Modeling Error Propagation as State Transitions in Multi-Agent LLM Pipelines — arXiv:2608.14588

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.14588
License (as archived): CC0-1.0
Archived: 2026-10-18T17:05:55.126Z
SHA-256 of the archived text: cc2c7bf5358217d58abd5c39adb19615742742ed8d0427640d96556e152ced75
Stable URL of this page: https://metraipolis.com/documento/dc_7800596dca71

How to cite

The Hallucination Snowball: Modeling Error Propagation as State Transitions in Multi-Agent LLM Pipelines — arXiv:2608.14588 — arXiv (en), https://arxiv.org/abs/2608.14588 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7800596dca71 (Metraipolis archive dc_7800596dca71, sha256 cc2c7bf5358217d58abd5c39adb19615742742ed8d0427640d96556e152ced75)

BibTeX

@misc{dc7800596dca71,
  title = {The Hallucination Snowball: Modeling Error Propagation as State Transitions in Multi-Agent LLM Pipelines — arXiv:2608.14588},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.14588},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7800596dca71; sha256 cc2c7bf5358217d58abd5c39adb19615742742ed8d0427640d96556e152ced75},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.14588. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Sequential multi-agent LLM pipelines chain specialized agents without verification at handoffs, creating a structural flaw with measurable and severe consequences. We show that hallucinations injected at Stage 1 do not merely persist; they transform: raw numerical facts become derived computations, then narrative prose, then editorially approved conclusions. At each transformation, detectability degrades near-irreversibly. We formalize this as the hallucination snowball effect, a first-order Markov process over four states (Raw Fact $\to$ Derived $\to$ Narrative $\to$ Invisible) with empirically measured per-boundary escape probabilities of 24.6%, 48.3%, and 89.3%. Across 346 automatically injected hallucinations in a 4-agent financial analysis pipeline on FinanceBench, gpt-4o detection drops from 72.0% at Stage 1 to 50.9% at Stage 4, and 23.7% of hallucinations survive completely undetected in the final output. Even the strongest model tested (Qwen3.5-397B-A17B, 87.0% at Stage 1) faces a structural ceiling; projected Stage 4 detection is only ${\sim}$60--65%. Critically, boundary gates using identical RAG verification tools reduce hallucination survival from 58.4% to 16.2% versus end-of-pipeline checking (Cohen's $h = -0.911$, $p < 0.000001$), while end-checking alone achieves merely 2.3 pp improvement over no verification. When you verify matters more than whether you verify. Our model predicts survival for $n$-agent linear pipelines and prescribes optimal verification resource allocation: invest at $S_1{\to}S_2$ first, where 75.4% of hallucinations are still catchable, not at $S_3{\to}S_4$ where 89.3% have already escaped.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-18 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2608.14588 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-18T16:29:13.708Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).