paperarXiv (systems, data, ML)·22 Oct 2026·Yancheng Zhu

When LLMs Forget What They Believe: Characterizing State Integrity Collapse in Long-Horizon Dialogues

Characterizes how LLMs keep repeating instructions while ceasing to obey them in long dialogues.

Read original on arxiv.org Read PDF Discussion

Depth 4/5Score 4.23

Paper (from PDF)

When LLMs Forget What They Believe: Characterizing State Integrity

Collapse in Long-Horizon Dialogues
YanchengZhu YifanYuan*
SchoolofArtificialIntelligence,ShenzhenUniversity,Shenzhen,China
2410673009@mails.szu.edu.cn,yifanyuan@szu.edu.cn
Abstract We call this failure State Integrity Collapse
--- -------- --- --- --- ------ ----------- ------- -------- ----------- --- --------
(SIC). SIC differs from ordinary forgetting. For-
Largelanguagemodelscanrepeataninstruc-
6202 tcO 22 ]LC.sc[ 1v41313.0162:viXra getting means the model cannot recover a prior
tionwhilefailingtokeepobeyingit. Westudy
-------------------------------- --- --- --- ------- --- --- --- --- --- --- ---
this recall–compliance dissociation in long- constraint;SICmeansthemodelrecoversitunder
horizondialogueanddefineStateIntegrityCol- probebutfailstomakeitgovernthenextanswer.
lapse: a model preserves a stated constraint The distinction matters for mitigation. Retrieval,
inmemorybutstopsapplyingittolaterturns.
memorysummarization,orlongercontextwindows
We introduce STATETRACK-BENCH, a con- may help when state is inaccessible, but they do
trolledsuiteofconversationscoveringnumeri-
notdirectlysolvefailuresofstateapplication.
cal,format,role/persona,factual-memory,and
The practical stakes are high. Consider a med-
------------------------ --- -------------------- --- --- --- --------- ------ --- -------------- --- ------
logical-ruleconstraints. Acrosssevenprovider-
pinned model snapshots and five dialogue- ical assistant instructed to avoid recommending
medicationscontraindicatedforapatient’scondi-
depthprobes,strictjointretentionrangesfrom
62.4% to 100.0%, with degradation concen- tion,oralegalassistantboundbyaconfidentiality
tratedinsymbolicconstraintsandinoneespe-
constraint. …

		Collapse		in Long-Horizon		Dialogues
				YanchengZhu	YifanYuan*
SchoolofArtificialIntelligence,ShenzhenUniversity,Shenzhen,China
2410673009@mails.szu.edu.cn,yifanyuan@szu.edu.cn
	Abstract				We	call this	failure	State	Integrity		Collapse
---	--------	---	---	---	------	-----------	-------	--------	-----------	---	--------
					(SIC).	SIC differs	from	ordinary	forgetting.		For-
Largelanguagemodelscanrepeataninstruc-
6202 tcO 22 ]LC.sc[ 1v41313.0162:viXra getting means the model cannot recover a prior
tionwhilefailingtokeepobeyingit.				Westudy
--------------------------------	---	---	---	-------	---	---	---	---	---	---	---
this recall–compliance dissociation in long- constraint;SICmeansthemodelrecoversitunder
horizondialogueanddefineStateIntegrityCol- probebutfailstomakeitgovernthenextanswer.
lapse: a model preserves a stated constraint The distinction matters for mitigation. Retrieval,
inmemorybutstopsapplyingittolaterturns.
memorysummarization,orlongercontextwindows
We introduce STATETRACK-BENCH, a con- may help when state is inaccessible, but they do
trolledsuiteofconversationscoveringnumeri-
notdirectlysolvefailuresofstateapplication.
cal,format,role/persona,factual-memory,and
					The	practical	stakes	are	high. Consider		a med-
------------------------	---	--------------------	---	---	---	---------	------	---	--------------	---	------
logical-ruleconstraints.		Acrosssevenprovider-
pinned model snapshots and five dialogue- ical assistant instructed to avoid recommending
medicationscontraindicatedforapatient’scondi-
depthprobes,strictjointretentionrangesfrom
62.4% to 100.0%, with degradation concen- tion,oralegalassistantboundbyaconfidentiality
tratedinsymbolicconstraintsandinoneespe-
					constraint.	…