# Lifting Manifolds to Mitigate Pseudo-Alignment in LLM4TS

arXiv: https://arxiv.org/abs/2510.12847 · HTML: https://mlsanity.com/paper/2510.12847/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 37 references that are in the map.
- First posted on arXiv 14 Oct 2025.

## Builds on (most cited first, 21 of 21)

- [A Time Series is Worth 64 Words: Long-term Forecasting with Transformers](https://mlsanity.com/paper/2211.14730/) (2022)
- [Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecasting](https://mlsanity.com/paper/2106.13008/) (2021)
- [Are Transformers Effective for Time Series Forecasting?](https://mlsanity.com/paper/2205.13504/) (2022)
- [FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecasting](https://mlsanity.com/paper/2201.12740/) (2022)
- [One Fits All:Power General Time Series Analysis by Pretrained LM](https://mlsanity.com/paper/2302.11939/) (2023)
- [Mind the Gap: Understanding the Modality Gap in Multi-modal Contrastive Representation Learning](https://mlsanity.com/paper/2203.02053/) (2022)
- [Large Language Models Are Zero-Shot Time Series Forecasters](https://mlsanity.com/paper/2310.07820/) (2023)
- [InceptionTime: Finding AlexNet for Time Series Classification](https://mlsanity.com/paper/1909.04939/) (2019)
- [Representation Degeneration Problem in Training Natural Language Generation Models](https://mlsanity.com/paper/1907.12009/) (2019)
- [A Multi-Horizon Quantile Recurrent Forecaster](https://mlsanity.com/paper/1711.11053/) (2017)
- [Are Language Models Actually Useful for Time Series Forecasting?](https://mlsanity.com/paper/2406.16964/) (2024)
- [UniTime: A Language-Empowered Unified Model for Cross-Domain Time Series Forecasting](https://mlsanity.com/paper/2310.09751/) (2023)
- [Adversarial Reprogramming of Neural Networks](https://mlsanity.com/paper/1806.11146/) (2018)
- [CALF: Aligning LLMs for Time Series Forecasting via Cross-modal Fine-Tuning](https://mlsanity.com/paper/2403.07300/) (2024)
- [TimeCMA: Towards LLM-Empowered Multivariate Time Series Forecasting via Cross-Modality Alignment](https://mlsanity.com/paper/2406.01638/) (2024)
- [Context-Alignment: Activating and Enhancing LLM Capabilities in Time Series](https://mlsanity.com/paper/2501.03747/) (2025)
- [Implicit Multimodal Alignment: On the Generalization of Frozen LLMs to Multimodal Inputs](https://mlsanity.com/paper/2405.16700/) (2024)
- [Adversarial Reprogramming Revisited](https://mlsanity.com/paper/2206.03466/) (2022)
- [Understanding Why Large Language Models Can Be Ineffective in Time Series Analysis: The Impact of Modality Alignment](https://mlsanity.com/paper/2410.12326/) (2024)
- [HydraRAG: Structured Cross-Source Enhanced Large Language Model Reasoning](https://mlsanity.com/paper/2505.17464/) (2025)
- [Irregularity-Informed Time Series Analysis: Adaptive Modelling of Spatial and Temporal Dynamics](https://mlsanity.com/paper/2410.12257/) (2024)

## Built on by (most cited first, 2 of 2)

- [MemoTime: Memory-Augmented Temporal Knowledge Graph Enhanced Large Language Model Reasoning](https://mlsanity.com/paper/2510.13614/) (2025)
- [Rethinking the Role of LLMs in Time Series Forecasting](https://mlsanity.com/paper/2602.14744/) (2026)

## Abstract (arXiv)

Pseudo-Alignment is a pervasive challenge in many large language models for time series (LLM4TS) models, often causing them to underperform compared to linear models or randomly initialised backbones. However, there is limited discussion in the community for the reasons that pseudo-alignment occurs. In this work, we conduct a thorough investigation into the root causes of pseudo-alignment in LLM4TS and build a connection of pseudo-alignment to the cone effect in LLM. We demonstrate that pseudo-alignment arises from the interplay of cone effect within pretrained LLM components and the intrinsically low-dimensional manifold of time-series data. In addition, we also introduce textitTimeSUP, a novel technique designed to mitigate this issue and improve forecast performance in existing LLM4TS approaches. TimeSUP addresses this by increasing the time series manifold to more closely match the intrinsic dimension of language embeddings, allowing the model to distinguish temporal signals clearly while still capturing shared structures across modalities. As a result, representations for time and language tokens remain distinct yet exhibit high cosine similarity, signifying that the model preserves each modality unique features while learning their commonalities in a unified embedding space. Empirically, TimeSUP consistently outperforms state-of-the-art LLM4TS methods and other lightweight baselines on long-term forecasting performance. Furthermore, it can be seamlessly integrated into four existing LLM4TS pipelines and delivers significant improvements in forecasting performance.