A methodical investigation of temporal prompting techniques for LLM-based EL is presented, and it is demonstrated that explicit temporal prompting can reduce drift mistakes by up to 40% using a dataset of temporally-sensitive mentions linked with Wikidata snapshots.
A new benchmark, KIFI, is designed, which comprises 1032 carefully selected instances from the TRUE and ScreenEval datasets, with key information annotated, and it is shown that LLMs frequently fail to use the appropriate information to make correct decisions.
Xindi Guo, Zhen Xie, Patrick H. Chen· Annual International ACM SIG...· 0 citations
This work presents the first systematic study of inverse relation directionality in LLMs, using a benchmark consisting of 5,457 instances spanning 27 distinct inverse relation labels and reveals systematic asymmetries in inverse relation classification across LLMs.
This framework proposes a context-editing framework that performs selective abstraction over entities that appear in both the context and the question, establishing symbolic abstraction as a highly cost-efficient solution for ensuring context fidelity in LLMs.
Rounak Sharma, Debabrata Mahapatra, S. Saini· Annual International ACM SIG...· 0 citations
Effectively understanding and modeling the temporal aspects of user queries is crucial for Information Retrieval (IR) and Question Answering (QA), particularly in contexts that demand freshness, historical accuracy, or temporal reasoning. In this paper, we present a formal and comprehensive taxonomy for classifying natural language queries along four dimensions: (i) temporal understanding, (ii) reasoning type, (iii) time sensitivity, and (iv) trendiness. This framework captures a wide range of temporal intents, from static factual questions to dynamic, event-driven queries. The taxonomy is designed as a foundation for diagnostic evaluation of QA and IR systems by providing a systematic categorization of the temporal properties of queries. This structure enables practitioners to identify, isolate, and analyze system behaviors across diverse temporal scenarios, helping uncover specific failure modes related to temporal reasoning, sensitivity, data freshness, and relevance. We apply the taxonomy to classify queries from four public datasets: MS MARCO, FreshQA, RealtimeQA, and SituatedQA, revealing systematic gaps and underrepresented time-sensitive categories. As a diagnostic case study, we stratify the accuracy of three LLMs by temporal dimension, showing that the taxonomy exposes failure patterns that the aggregate metrics conceal.
Ivano Lauriola· International Conference on...· 0 citations
Information continuously evolves over time. Because of this dynamic nature, time becomes a fundamental dimension that shapes how we extract, retrieve, interpret, and reason about knowledge. As information systems are constantly updated, models must determine not only what is relevant, but also when that information is valid. This tutorial provides a structured and in-depth overview of the complete temporal information access pipeline: Temporal Information Extraction (TIE), Temporal Information Retrieval (TIR), and Temporal Question Answering (TQA). We examine the progression of temporal methods from early rule-based extraction and probabilistic retrieval to contemporary transformer-based and large language model (LLM) architectures. Participants gain a solid understanding of the core principles underlying the identification and normalization of time expressions, time-aware document ranking, and temporal reasoning in retrieval-augmented generation (RAG). The tutorial concludes with a discussion of open challenges and future research directions aimed at building AI systems that are temporally aware, robust, and adaptive. By connecting classical extraction and IR foundations with modern LLM-based reasoning, this tutorial presents a cohesive and up-to-date perspective on temporal information systems.
Bhawna Piryani, Avishek Anand, Omar Alonso et al.· Annual International ACM SIG...· 0 citations
Model editing keeps large language models (LLMs) up to date without retraining, but temporal facts expose a limitation of the prevailing locate-and-edit paradigm: an update is not always a replacement. When a fact changes, the new answer should become current while the old answer may remain correct in historical time contexts. Building on this insight, we use causal tracing to show that LLMs already support this distinction via a two-stage internal computation: early MLP layers retrieve a time-agnostic subject representation, and later layers modulate it with temporal context to yield the time-correct answer. Motivated by this finding, we introduce PRISM Edit, which optimizes a single polysemous representation across temporal contexts and leverages the model's inherent modulation pathway to route it to temporally correct predictions without requiring any architectural modification. We evaluate on TimeConflict, a newly introduced temporal editing benchmark, and on temporally augmented CounterFact. PRISM Edit improves multiple core metrics over the best baseline, most notably +23.3 Temporal Consistency (TC) and +33.7 Current Relative-time Score (CRS) on LLaMA-3, while being more than 2x faster. Code and data are publicly available at https://github.com/CheerCHuang/PRISM-Edit.
Chen Huang, Qirui Zheng, Ruiqin Zheng et al.· 1 citation