arXiv：cs.AI· Yongyi Huang, Huaping Zhang, Wenyao Cui, Qiuchi Li, Lei Li, Baohua Zhang·· 17 小时前
ChronoRegime：评估 LLM 未来可预测性的时间推理新任务与 CTR 框架
Is the Future Predictable? Modeling Future Predictability for Reliable Temporal Reasoning
AI 导读
研究者提出 ChronoRegime，一个面向未来的时间推理任务，要求模型判断历史时间证据是否足以支撑对某一未来答案下结论，并设计 Confidence-Aware Temporal Reasoning（CTR）框架，从时间动态中估计未来可预测性并据此调控作答。
正文
Abstract:Large language models (LLMs) remain unreliable when reasoning about the future. Existing models often overlook the difference among future states, committing to specific answers even when the queried facts are inherently volatile. To capture this challenge, we introduce ChronoRegime, a future-oriented temporal reasoning task that requires models to determine whether historical temporal evidence justifies commitment to a future answer. Furthermore, we propose Confidence-Aware Temporal Reasoning (CTR), a framework that estimates future predictability from temporal dynamics and regulates answer commitment accordingly. Rather than performing unconditional forecasting, CTR commits only when the reconstructed evidence supports the target future state. Across multiple LLMs, CTR improves strict overall accuracy by 6.7 percentage points and reduces hallucination rates by 18.4 percentage points on average over the plain QA baseline. These results demonstrate that explicit predictability assessment is an effective mechanism for safer and more reliable temporal reasoning under future uncertainty.
| Comments: | Accepted to EMNLP 2026 (Main Conference) |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.14073 [cs.CL] |
| (or arXiv:2610.14073v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.14073 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Baohua Zhang [view email]
[v1]
Thu, 8 Oct 2026 03:46:15 UTC (2,812 KB)
来源：arXiv：cs.AI · arxiv.org