# Time Present and Time Past: Benchmarking Large Language Models on Temporally Evolving Document Understanding

arXiv: https://arxiv.org/abs/2608.08512 · HTML: https://mlsanity.com/paper/2608.08512/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 48 references that are in the map.
- First posted on arXiv 9 Aug 2026.

## Builds on (most cited first, 18 of 18)

- [gpt-oss-120b & gpt-oss-20b Model Card](https://mlsanity.com/paper/2508.10925/) (2025)
- [DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence](https://mlsanity.com/paper/2606.19348/) (2026)
- [In-Context Retrieval-Augmented Language Models](https://mlsanity.com/paper/2302.00083/) (2023)
- [GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models](https://mlsanity.com/paper/2508.06471/) (2025)
- [MKQA: A Linguistically Diverse Benchmark for Multilingual Open Domain Question Answering](https://mlsanity.com/paper/2007.15207/) (2020)
- [StreamingQA: A Benchmark for Adaptation to New Knowledge over Time in Question Answering Models](https://mlsanity.com/paper/2205.11388/) (2022)
- [XOR QA: Cross-lingual Open-Retrieval Question Answering](https://mlsanity.com/paper/2010.11856/) (2020)
- [TemporalWiki: A Lifelong Benchmark for Training and Evaluating Ever-Evolving Language Models](https://mlsanity.com/paper/2204.14211/) (2022)
- [Test of Time: A Benchmark for Evaluating LLMs on Temporal Reasoning](https://mlsanity.com/paper/2406.09170/) (2024)
- [UnSeenTimeQA: Time-Sensitive Question-Answering Beyond LLMs' Memorization](https://mlsanity.com/paper/2407.03525/) (2024)
- [MRAG: A Modular Retrieval Framework for Time-Sensitive Question Answering](https://mlsanity.com/paper/2412.15540/) (2024)
- [EvoWiki: Evaluating LLMs on Evolving Knowledge](https://mlsanity.com/paper/2412.13582/) (2024)
- [ComplexTempQA:A 100m Dataset for Complex Temporal Question Answering](https://mlsanity.com/paper/2406.04866/) (2024)
- [When Benchmarks Age: Temporal Misalignment through Large Language Model Factuality Evaluation](https://mlsanity.com/paper/2510.07238/) (2025)
- [VersionRAG: Version-Aware Retrieval-Augmented Generation for Evolving Documents](https://mlsanity.com/paper/2510.08109/) (2025)
- [MathMist: A Parallel Multilingual Benchmark Dataset for Mathematical Problem Solving and Reasoning](https://mlsanity.com/paper/2510.14305/) (2025)
- [Can Large Language Models Keep Up? Benchmarking Online Adaptation to Continual Knowledge Streams](https://mlsanity.com/paper/2603.07392/) (2026)
- [When Facts Change: Probing LLMs on Evolving Knowledge with evolveQA](https://mlsanity.com/paper/2510.19172/) (2025)

## Built on by (most cited first, 1 of 1)

- [Authority Before Utility: Non-Compensatory Control for Persistent LLM Memory](https://mlsanity.com/paper/2609.37474/) (2026)

## Abstract (arXiv)

Evolving documents, such as laws, tax codes, and software documentation, are amended, replaced, and sometimes reverted over time, so a question has different correct answers at different dates. In contrast to encyclopedic knowledge, where an old fact is simply overwritten, an amendment is itself an official text that states what it replaces and when it takes effect, and the earlier version stays correct for its validity period. The central challenge is therefore version resolution, that is, identifying the version in force on the queried date. Existing temporal QA datasets treat time only as an annotation, so version resolution stays untested. We present TIDE, an expert-verified benchmark of 3,050 QA pairs over 644 official customs instruments issued between 1969 and 2025 by the Government of Bangladesh, covering eight task types over deeply code-mixed documents that are heterogeneous in layout and dated in two calendars. In addition, we evaluate nine recent LLMs under a single protocol across parametric, gold-context, and retrieval access, scored by a three-judge LLM council with a hard date gate separating correct meaning from correct time. The best macro-averaged accuracy is only 68.5%. Resolving a version from an implicit date reaches 59.7%, and detecting that the supplied version does not govern the query reaches only 26.7%. Models are more likely to find correct versions than to reject incorrect ones, and they tend to follow a confident parametric answer over the supplied authoritative text. All code and data are available at https://github.com/icsetepa44/TIDE