Pith. sign in

REVIEW 14 cited by

Large Language Models Are Zero-Shot Time Series Forecasters

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2310.07820 v3 pith:43A67ZYO submitted 2023-10-11 cs.LG

Large Language Models Are Zero-Shot Time Series Forecasters

classification cs.LG
keywords seriestimellmsmodelsperformancedatadistributionsfind
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

By encoding time series as a string of numerical digits, we can frame time series forecasting as next-token prediction in text. Developing this approach, we find that large language models (LLMs) such as GPT-3 and LLaMA-2 can surprisingly zero-shot extrapolate time series at a level comparable to or exceeding the performance of purpose-built time series models trained on the downstream tasks. To facilitate this performance, we propose procedures for effectively tokenizing time series data and converting discrete distributions over tokens into highly flexible densities over continuous values. We argue the success of LLMs for time series stems from their ability to naturally represent multimodal distributions, in conjunction with biases for simplicity, and repetition, which align with the salient features in many time series, such as repeated seasonal trends. We also show how LLMs can naturally handle missing data without imputation through non-numerical text, accommodate textual side information, and answer questions to help explain predictions. While we find that increasing model size generally improves performance on time series, we show GPT-4 can perform worse than GPT-3 because of how it tokenizes numbers, and poor uncertainty calibration, which is likely the result of alignment interventions such as RLHF.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 14 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. WorldReasoner: Evaluating Whether Language Model Agents Forecast Events with Valid Reasoning

    cs.CL 2026-06 unverdicted novelty 7.0

    WorldReasoner supplies 345 resolved forecasting tasks built from 14,141 articles to score LM agents on outcome quality, evidence quality, and reasoning quality against time-bounded evidence and hindsight graphs.

  2. TSVer: A Benchmark for Fact Verification Against Time-Series Evidence

    cs.CL 2025-11 unverdicted novelty 7.0

    TSVer is a new benchmark dataset for fact verification against time-series evidence, with 304 annotated real-world claims, 400 time series, verdicts, and justifications, plus baseline results showing current models struggle.

  3. Pre-trained Large Language Models Learn Hidden Markov Models In-context

    cs.LG 2025-06 unverdicted novelty 7.0

    Pre-trained LLMs learn to predict HMM-generated sequences via in-context learning, approaching theoretical optimum on synthetic HMMs and matching expert models on real animal decision data.

  4. Sundial: A Family of Highly Capable Time Series Foundation Models

    cs.LG 2025-02 conditional novelty 7.0

    Sundial uses TimeFlow Loss for native pre-training of Transformers on continuous time series from TimeBench, achieving SOTA point and probabilistic forecasting with millisecond inference.

  5. A decoder-only foundation model for time-series forecasting

    cs.CL 2023-10 unverdicted novelty 7.0

    A pretrained decoder-only patched transformer achieves near state-of-the-art zero-shot forecasting performance across diverse time series datasets and settings.

  6. LLM-Guided Task-Semantic Field Factorization for Industrial Process Forecasting

    cs.LG 2026-07 conditional novelty 6.0

    Offline LLM-built variable-semantic directions, activated by each numerical window via a dual-path adapter, cut industrial forecasting MAE up to 25.5% with ~2–3k extra parameters.

  7. ReflectiChain: Epistemic Grounding in LLM-Driven World Models for Supply Chain Resilience

    cs.AI 2026-06 unverdicted novelty 6.0

    A framework combining LLM policy interpretation with a physically conserved graph-latent world model and uncertainty-separated learning achieves 33% higher rationale consistency and 82.3% operability on a 10-node semi...

  8. GITCO: Gated Inference-Time Context Optimization in TSFMs

    cs.AI 2026-06 unverdicted novelty 6.0

    GITCO delivers +1.95% average MASE reduction on TimesFM 2.5 across 53 datasets by gated inference-time suppression of anomalous patches, capturing 89.9% of the improvement upper bound.

  9. VESTA: Visual Exploration with Statistical Tool Agents

    cs.AI 2026-05 unverdicted novelty 6.0

    VESTA introduces dynamic tool creation for VLMs that outperforms static-tool and no-tool baselines on distribution fitting, time series, and astronomy tasks in the new DAWN benchmark.

  10. InfoAtlas: A Foundation Model for Zero-Shot Statistical Dependence Estimate

    cs.LG 2026-05 unverdicted novelty 6.0

    InfoAtlas is a pretrained neural model for zero-shot mutual information estimation that matches state-of-the-art accuracy with 100x speedup and handles varying dimensions via a single model.

  11. E4GEN: Event-level Explainable Extreme-Enhanced Time-series Generation

    cs.LG 2026-06 unverdicted novelty 5.0

    E4GEN is an explainable diffusion model using E-Activator, E-Predictor, and E-Control for extreme-event-aware time-series generation evaluated on six datasets.

  12. Degradation-aware Predictive Energy Management for Fuel Cell-Battery Ship Power System with Data-driven Load Forecasting

    eess.SY 2026-04 unverdicted novelty 5.0

    A degradation-aware predictive controller for hybrid ship power systems reduces hydrogen consumption by up to 5.8% and fuel cell degradation by up to 36.4% versus a filter-based benchmark on real harbor tug data.

  13. AR-KAN: Autoregressive-Weight-Enhanced Kolmogorov-Arnold Network for Time Series Forecasting

    cs.LG 2025-09 unverdicted novelty 5.0

    AR-KAN combines a pre-trained AR module with KAN to reduce redundancy while preserving temporal features, delivering lower probabilistic approximation error and stronger forecasting results on synthetic almost-periodi...

  14. LLM-Guided Task-Semantic Field Factorization for Industrial Process Forecasting

    cs.LG 2026-07 conditional novelty 4.0

    TSF converts process variable documents into frozen semantic vectors that scale the numerical input window before a time-series backbone, yielding average MAE reductions of 2.9–3.6% across industrial forecasting tasks.