Metraipolis › archive › dc_2bb25f8fe53c
When LLMs Forget What They Believe: Characterizing State Integrity Collapse in Long-Horizon Dialogues — arXiv:2610.31314
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.31314
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-23T15:15:26.256Z
- SHA-256 of the archived text
1bae1bf9d2ef40b86428c0544b184de03923c01f911859bf520c8b9178407d70- Stable URL of this page
https://metraipolis.com/documento/dc_2bb25f8fe53c
How to cite
When LLMs Forget What They Believe: Characterizing State Integrity Collapse in Long-Horizon Dialogues — arXiv:2610.31314 — arXiv (en), https://arxiv.org/abs/2610.31314 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2bb25f8fe53c (Metraipolis archive dc_2bb25f8fe53c, sha256 1bae1bf9d2ef40b86428c0544b184de03923c01f911859bf520c8b9178407d70)
BibTeX
@misc{dc2bb25f8fe53c,
title = {When LLMs Forget What They Believe: Characterizing State Integrity Collapse in Long-Horizon Dialogues — arXiv:2610.31314},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.31314},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2bb25f8fe53c; sha256 1bae1bf9d2ef40b86428c0544b184de03923c01f911859bf520c8b9178407d70},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.31314. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models can repeat an instruction while failing to keep obeying it. We study this recall-compliance dissociation in long-horizon dialogue and define State Integrity Collapse: a model preserves a stated constraint in memory but stops applying it to later turns. We introduce StateTrack-Bench, a controlled suite of conversations covering numerical, format, role/persona, factual-memory, and logical-rule constraints. Across seven provider-pinned model snapshots and five dialogue-depth probes, strict joint retention ranges from 62.4% to 100.0%, with degradation concentrated in symbolic constraints and in one especially brittle model. A forked recall audit shows that 88.6% of observed failures are behavioral-drift failures rather than ordinary forgetting; excluding the brittle outlier preserves this diagnosis (84.4% of failures). Extended evaluations at depths up to 50 turns confirm that the dissociation persists beyond the core benchmark, while Structured State Anchoring recovers up to 24.4 points on immutable constraints. The results show that reliability cannot be inferred from single-turn instruction following alone; evaluation must track whether remembered constraints remain operational as dialogue state grows. Code and artifacts are available at: this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.31314 — 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 2026-10-23T15:09:08.607Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).