Time to Revist Exact Match
On the map: LLM reasoning → Knowledge- and tool-augmented reasoning
Abstract (arXiv)
Temporal question answering is an established method for evaluating temporal reasoning in large language models. Expected answers are often numeric (e.g., dates or durations), yet model responses are evaluated like regular text with exact match (EM), unable to distinguish small from large errors. In this investigative work, we frame temporal question answering as a numerical estimation task to assess the shortcomings of EM. We introduce TempAnswerQA, a benchmark distilled from Test of Time and TempTabQA, where all questions require a numerical, temporal answer, allowing us to evaluate models beyond EM. We use the forecasting metrics symmetric mean absolute percentage error (sMAPE) and mean absolute scaled error (MASE). With sMAPE, we find that error size and EM are decoupled. Models with low EM still have low sMAPE (both ~20%), and some models have high sMAPE despite high EM. Scaling errors by the deviation of the ground truth data with MASE reshuffles model rankings compared to EM, revealing gaps in models' understanding of temporal domain knowledge, especially when trained with synthetic data. Lastly, the models' most frequent error is to deviate by only ±1 from the ground truth. sMAPE and MASE, unlike EM, adequately weight these errors. Our findings underscore the need for specialised metrics for temporal QA tasks. Code and data are available on https://github.com/aauss/temporal-answer-qa.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Knowledge- and tool-augmented reasoning, one of 17 directions in LLM reasoning. Builds on 13 of its 29 references that are in the map. First posted on arXiv 20 Sep 2025.
Builds on
13 of 29 references.
Closest ancestors · 4
Same direction, newest first
- 2024Enhancing Temporal Understanding in LLMs for Semi-structured Tables22 Jul
- 2024Timo: Towards Better Temporal Reasoning for Language Models20 Jun
- 2024Test of Time: A Benchmark for Evaluating LLMs on Temporal Reasoning13 Jun
- 2024Towards Effective Time-Aware Language Representation: Exploring Enhanced Temporal Understanding in Language Models4 Jun
Other references · 9
- 2019BERTScore: Evaluating Text Generation with BERTcited by 5,512
- 2023BloombergGPT: A Large Language Model for Financecited by 939
- 2024Libra: Leveraging Temporal Images for Biomedical Radiology Analysiscited by 29
- 2023TempTabQA: Temporal Question Answering for Semi-Structured Tablescited by 27
- 2021A better measure of relative prediction accuracy for model selection and model estimationcited by 20
- 2025ChronoSense: Exploring Temporal Understanding in Large Language Models with Time Intervals of Eventscited by 18
- 2025A Study into Investigating Temporal Robustness of LLMscited by 11
- 2024Temporal Knowledge Graph Question Answering: A Surveycited by 9
- 2025Factual Knowledge in Language Models: Robustness and Anomalies under Simple Temporal Context Variationscited by 1
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2024Set the Clock: Temporal Alignment of Pretrained Language Models
- 2025It's High Time: A Survey of Temporal Question Answering
- 2025TRAVELER: A Benchmark for Evaluating Temporal Reasoning across Vague, Implicit and Explicit References
- 2024Time Awareness in Large Language Models: Benchmarking Fact Recall Across Time
- 2023Unlocking Temporal Question Answering for Large Language Models with Tailor-Made Reasoning Logic
- 2023Towards Benchmarking and Improving the Temporal Reasoning Capability of Large Language Models
- 2023MenatQA: A New Dataset for Testing the Temporal Comprehension and Reasoning Abilities of Large Language Models
- 2026AdapTime: Enabling Adaptive Temporal Reasoning in Large Language Models