Evaluating Temporal Consistency in Multi-Turn Language Models

2026-04-24

AI Summary

This work addresses the challenge that language models often fail to maintain implicit temporal assumptions in multi-turn dialogues due to the absence of explicit temporal references, leading to temporal inconsistencies. The authors introduce the notion of “temporal scope stability” and present ChronoScope, the first million-scale, controllable multi-turn dialogue benchmark grounded in Wikidata. ChronoScope enables systematic evaluation of models’ capabilities in preserving, switching, and transferring temporal context through deterministic generation, knowledge grounding, and fine-grained dialogue control. Experimental results reveal a pervasive “temporal drift” phenomenon among mainstream models: even when equipped with correct factual knowledge, they tend to default to a present-time assumption, and such errors compound as dialogue progresses.

Technology Category

• Natural Language Processing: Language Grounding & Multi-modal NLP
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Temporal Planning

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Language models are increasingly deployed in interactive settings where users reason about facts over time rather than in isolation. In such scenarios, correct behavior requires models to maintain and update implicit temporal assumptions established earlier in a conversation. We study this challenge through the lens of temporal scope stability: the ability to preserve, override, or transfer time-scoped factual context across dialogue turns. We introduce ChronoScope, a large-scale diagnostic benchmark designed to isolate temporal scope behavior in controlled multi-turn interactions, comprising over one million deterministically generated question chains grounded in Wikidata. ChronoScope evaluates whether models can correctly retain inferred temporal scope when follow-up questions omit explicit time references, spanning implicit carryover, explicit scope switching, cross-entity transfer, and longer temporal trajectories. Through extensive evaluation of state-of-the-art language models, we find that temporal scope stability is frequently violated in controlled multi-turn settings, with models often drifting toward present-day assumptions despite correct underlying knowledge. These failures intensify with interaction length and persist even under oracle context conditions, revealing a gap between single-turn factual accuracy and coherent temporal reasoning under sequential interaction. We make our dataset and evaluation suite publicly available at https://github.com/yashkumaratri/ChronoScope

Problem

• temporal consistency
• multi-turn dialogue
• temporal scope
• language models
• temporal reasoning

Innovation

• temporal scope stability
• multi-turn language models
• ChronoScope
• temporal consistency
• time-scoped reasoning