# Unveiling Divergent Inductive Biases of LLMs on Temporal Data

arXiv: https://arxiv.org/abs/2404.01453 · HTML: https://mlsanity.com/paper/2404.01453/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 16 references that are in the map.
- First posted on arXiv 1 Apr 2024.

## Builds on (most cited first, 9 of 9)

- [A Multi-Axis Annotation Scheme for Event Temporal Relations](https://mlsanity.com/paper/1804.07828/) (2018)
- [Temporal Reasoning on Implicit Events from Distant Supervision](https://mlsanity.com/paper/2010.12753/) (2020)
- [Joint Constrained Learning for Event-Event Relation Extraction](https://mlsanity.com/paper/2010.06727/) (2020)
- [An Improved Neural Baseline for Temporal Relation Extraction](https://mlsanity.com/paper/1909.00429/) (2019)
- [CogCompTime: A Tool for Understanding Time in Natural Language Text](https://mlsanity.com/paper/1906.04940/) (2019)
- [Extracting Temporal Event Relation with Syntax-guided Graph Transformer](https://mlsanity.com/paper/2104.09570/) (2021)
- [A Survey on Event-based News Narrative Extraction](https://mlsanity.com/paper/2302.08351/) (2023)
- [A Sequential Model for Classifying Temporal Relations between Intra-Sentence Events](https://mlsanity.com/paper/1707.07343/) (2017)
- [What Makes the Story Forward? Inferring Commonsense Explanations as Prompts for Future Event Generation](https://mlsanity.com/paper/2201.07099/) (2022)

## Built on by (most cited first, 4 of 4)

- [Reliable Post-Retrieval Assembly for Agent Memory: Separating Evidence Extraction from Policy Execution](https://mlsanity.com/paper/2606.01435/) (2026)
- [EventFull: Complete and Consistent Event Relation Annotation](https://mlsanity.com/paper/2412.12733/) (2024)
- [DateLogicQA: Benchmarking Temporal Biases in Large Language Models](https://mlsanity.com/paper/2412.13377/) (2024)
- [Correct-Detect: Balancing Performance and Ambiguity Through the Lens of Coreference Resolution in LLMs](https://mlsanity.com/paper/2509.14456/) (2025)

## Abstract (arXiv)

Unraveling the intricate details of events in natural language necessitates a subtle understanding of temporal dynamics. Despite the adeptness of Large Language Models (LLMs) in discerning patterns and relationships from data, their inherent comprehension of temporal dynamics remains a formidable challenge. This research meticulously explores these intrinsic challenges within LLMs, with a specific emphasis on evaluating the performance of GPT-3.5 and GPT-4 models in the analysis of temporal data. Employing two distinct prompt types, namely Question Answering (QA) format and Textual Entailment (TE) format, our analysis probes into both implicit and explicit events. The findings underscore noteworthy trends, revealing disparities in the performance of GPT-3.5 and GPT-4. Notably, biases toward specific temporal relationships come to light, with GPT-3.5 demonstrating a preference for "AFTER'' in the QA format for both implicit and explicit events, while GPT-4 leans towards "BEFORE''. Furthermore, a consistent pattern surfaces wherein GPT-3.5 tends towards "TRUE'', and GPT-4 exhibits a preference for "FALSE'' in the TE format for both implicit and explicit events. This persistent discrepancy between GPT-3.5 and GPT-4 in handling temporal data highlights the intricate nature of inductive bias in LLMs, suggesting that the evolution of these models may not merely mitigate bias but may introduce new layers of complexity.