# Rethinking the Role of LLMs in Time Series Forecasting

arXiv: https://arxiv.org/abs/2602.14744 · HTML: https://mlsanity.com/paper/2602.14744/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Time series reasoning, one of 17 directions in LLM reasoning.
- Builds on 26 of its 94 references that are in the map.
- First posted on arXiv 16 Feb 2026.

## Placed in

- LLM reasoning / [Time series reasoning](https://mlsanity.com/lr/time-series-reasoning/)

## Builds on (most cited first, 26 of 26)

- [Categorical Reparameterization with Gumbel-Softmax](https://mlsanity.com/paper/1611.01144/) (2016)
- [Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation](https://mlsanity.com/paper/1308.3432/) (2013)
- [CoCa: Contrastive Captioners are Image-Text Foundation Models](https://mlsanity.com/paper/2205.01917/) (2022)
- [Instruction-Following Evaluation for Large Language Models](https://mlsanity.com/paper/2311.07911/) (2023)
- [Chronos: Learning the Language of Time Series](https://mlsanity.com/paper/2403.07815/) (2024)
- [Time-LLM: Time Series Forecasting by Reprogramming Large Language Models](https://mlsanity.com/paper/2310.01728/) (2023)
- [Large Language Models for Time Series: A Survey](https://mlsanity.com/paper/2402.01801/) (2024)
- [TimeFilter: Patch-Specific Spatial-Temporal Graph Filtration for Time Series Forecasting](https://mlsanity.com/paper/2501.13041/) (2025)
- [Empowering Time Series Analysis with Large Language Models: A Survey](https://mlsanity.com/paper/2402.03182/) (2024)
- [TimeXL: Explainable Multi-modal Time Series Prediction with LLM-in-the-Loop](https://mlsanity.com/paper/2503.01013/) (2025)
- [Enhancing Foundation Models for Time Series Forecasting via Wavelet-based Tokenization](https://mlsanity.com/paper/2412.05244/) (2024)
- [Multi-Patch Prediction: Adapting LLMs for Time Series Representation Learning](https://mlsanity.com/paper/2402.04852/) (2024)
- [Time-o1: Time-Series Forecasting Needs Transformed Label Alignment](https://mlsanity.com/paper/2505.17847/) (2025)
- [Visual question answering: from early developments to recent advances -- a survey](https://mlsanity.com/paper/2501.03939/) (2025)
- [From Values to Tokens: An LLM-Driven Framework for Context-aware Time Series Forecasting via Symbolic Discretization](https://mlsanity.com/paper/2508.09191/) (2025)
- [UMIT: Unifying Medical Imaging Tasks via Vision-Language Models](https://mlsanity.com/paper/2503.15892/) (2025)
- [Towards Cross-Modality Modeling for Time Series Analytics: A Survey in the LLM Era](https://mlsanity.com/paper/2505.02583/) (2025)
- [Relational Conformal Prediction for Correlated Time Series](https://mlsanity.com/paper/2502.09443/) (2025)
- [Using Pre-trained LLMs for Multivariate Time Series Forecasting](https://mlsanity.com/paper/2501.06386/) (2025)
- [LETS Forecast: Learning Embedology for Time Series Forecasting](https://mlsanity.com/paper/2506.06454/) (2025)
- [From Tokenizer Bias to Backbone Capability: A Controlled Study of LLMs for Time Series Forecasting](https://mlsanity.com/paper/2504.08818/) (2025)
- [The Few Govern the Many:Unveiling Few-Layer Dominance for Time Series Models](https://mlsanity.com/paper/2511.07237/) (2025)
- [Lifting Manifolds to Mitigate Pseudo-Alignment in LLM4TS](https://mlsanity.com/paper/2510.12847/) (2025)
- [The Solution for the CVPR2024 NICE Image Captioning Challenge](https://mlsanity.com/paper/2404.12739/) (2024)
- [Enhancing LLMs for Time Series Forecasting via Structure-Guided Cross-Modal Alignment](https://mlsanity.com/paper/2505.13175/) (2025)
- [TimeCF: A TimeMixer-Based Model with adaptive Convolution and Sharpness-Aware Minimization Frequency Domain Loss for long-term time seris forecasting](https://mlsanity.com/paper/2505.17532/) (2025)

## Built on by (most cited first, 3 of 3)

- [What Do Visual Tokens Really Encode? Uncovering Sparsity and Redundancy in Multimodal Large Language Models](https://mlsanity.com/paper/2603.00510/) (2026)
- [LLM Pretraining Shapes a Generalizable Manifold: Insights into Cross-Modal Transfer to Time Series](https://mlsanity.com/paper/2605.20449/) (2026)
- [From Recognition to Understanding: Unlocking Cognitive Time Series Reasoning with LLMs](https://mlsanity.com/paper/2606.22126/) (2026)

## Abstract (arXiv)

Large language models (LLMs) have been introduced to time series forecasting (TSF) to incorporate contextual knowledge beyond numerical signals. However, existing studies question whether LLMs provide genuine benefits, often reporting comparable performance without LLMs. We show that such conclusions stem from limited evaluation settings and do not hold at scale. We conduct a large-scale study of LLM-based TSF (LLM4TSF) across 8 billion observations, 17 forecasting scenarios, 4 horizons, multiple alignment strategies, and both in-domain and out-of-domain settings. Our results demonstrate that LLM4TS indeed improves forecasting performance, with especially large gains in cross-domain generalization. Pre-alignment outperforming post-alignment in over 90% of tasks. Both pretrained knowledge and model architecture of LLMs contribute and play complementary roles: pretraining is critical under distribution shifts, while architecture excels at modeling complex temporal dynamics. Moreover, under large-scale mixed distributions, a fully intact LLM becomes indispensable, as confirmed by token-level routing analysis and prompt-based improvements. Overall, Our findings overturn prior negative assessments, establish clear conditions under which LLMs are not only useful, and provide practical guidance for effective model design. We release our code at https://github.com/EIT-NLP/LLM4TSF.