Metraipolis › archive › dc_786778cd3e7e
Persistent Semantic Entities in Tool-Augmented LLM Systems — arXiv:2608.07952
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.07952
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T08:58:29.858Z
- SHA-256 of the archived text
d98d11ceed2fa784cab804e033d9d04aa0d95e26e5805c0a87a4b382880c0c73- Stable URL of this page
https://metraipolis.com/documento/dc_786778cd3e7e
How to cite
Persistent Semantic Entities in Tool-Augmented LLM Systems — arXiv:2608.07952 — arXiv (en), https://arxiv.org/abs/2608.07952 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_786778cd3e7e (Metraipolis archive dc_786778cd3e7e, sha256 d98d11ceed2fa784cab804e033d9d04aa0d95e26e5805c0a87a4b382880c0c73)
BibTeX
@misc{dc786778cd3e7e,
title = {Persistent Semantic Entities in Tool-Augmented LLM Systems — arXiv:2608.07952},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.07952},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_786778cd3e7e; sha256 d98d11ceed2fa784cab804e033d9d04aa0d95e26e5805c0a87a4b382880c0c73},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.07952. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Tool-augmented LLM agents can harbor implicit state that persists across sessions, activates through events, and propagates across agent boundaries---largely invisible to standard debugging. We formalize this as Persistent Semantic Entities (PSEs): constructs defined by name binding, event triggering, and cross-boundary propagation, and evaluate them across 24 models from 11 families (1.5B--1T parameters). First, every tested model is susceptible (20--100% on the 20-model susceptibility panel), with name binding as the necessary and dominant mechanism: without it, contamination is 0%. Second, persistence depends on contamination type rather than scale or deployment: preference contamination persists undecayed on every model probed (100% at t=10) and instruction contamination persists wherever adopted, persona-style injection decays partially (90%$\to$10%), while factual injection is model-dependent---self-corrected on Llama-3.1-8B and GPT-4o-mini but held at ceiling on both Qwen2.5-coder variants, so we do not claim it self-corrects in general. The preference and instruction results hold across providers in our controlled setting. Third, context-isolated self-verification achieves 20--79% reduction (median 36.5%) without oracle references while keyword-based detection produces systematic false positives, and contamination compounds 1.9$\times$ along a four-stage agent pipeline (40%$\to$75%). Preference and instruction contamination---persistent, lacking self-correction, and poorly captured by standard monitoring---represent a particularly concerning attack surface for deployed agent systems. Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-08-11 in cs.CR). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2608.07952 — 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 2026-10-19T08:54:24.988Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-08-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).