mlsanityData to 14 Oct 2026

T-SMART: Mechanism-Level Attribution for Tool-Augmented Time-Series Question Answering

arXiv 2609.14142 ↗First posted 12 Sep 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Large language models (LLMs) can struggle with time-series question answering (TS-QA), especially when numerical signals are serialized as text and require explicit computation. Tool-augmented approaches improve performance, but existing systems often intertwine language reasoning, computation, and perception, making it difficult to determine which components drive the gains. We present T-SMART, a neurosymbolic framework that separates these roles: a frozen LLM interprets questions and selects operations, deterministic tools perform numerical computation, and structured perception is invoked only when needed. Controlled paired ablations show that deterministic computation provides the dominant benefit, improving accuracy by 31.7 percentage points over direct LLM reasoning on serialized time series, while language understanding and perception offer smaller complementary gains. These results indicate that tool-augmented TS-QA benefits primarily from reliable numerical execution rather than additional language-model reasoning and provide a controlled framework for analyzing component contributions in neurosymbolic time-series systems.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 28 of its 33 references that are in the map. First posted on arXiv 12 Sep 2026.

Builds on

28 of 33 references.

References · 28

  1. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  2. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  3. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  4. 2022Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Taskscited by 1,076
  5. 2023Large Language Models Cannot Self-Correct Reasoning Yetcited by 1,017
  6. 2022PAL: Program-aided Language Modelscited by 652
  7. 2023Large Language Models Are Zero-Shot Time Series Forecasterscited by 500
  8. 2024To CoT or not to CoT? Chain-of-thought helps mainly on math and symbolic reasoningcited by 263
  9. 2024Are Language Models Actually Useful for Time Series Forecasting?cited by 168
  10. 2022DePlot: One-shot visual language reasoning by plot-to-table translationcited by 160
  11. 2024Language Models Still Struggle to Zero-shot Reason about Time Seriescited by 88
  12. 2025Time-MQA: Time Series Multi-Task Question Answering with Context Enhancementcited by 82
16 more
  1. 2025MTBench: A Multimodal Time Series Benchmark for Temporal Reasoning and Question Answeringcited by 55
  2. 2024TimeSeriesExam: A time series understanding examcited by 49
  3. 2026TimeART: Towards Agentic Time Series Reasoning via Tool-Augmentationcited by 26
  4. 2025A Survey of Reasoning and Agentic Systems in Time Series with Large Language Modelscited by 21
  5. 2026TSRBench: A Comprehensive Multi-task Multi-modal Time Series Reasoning Benchmark for Generalist Modelscited by 18
  6. 2024Sycophancy in Vision-Language Models: A Systematic Analysis and an Inference-Time Mitigation Frameworkcited by 17
  7. 2026TSAQA: Time Series Analysis Question And Answering Benchmarkcited by 11
  8. 2024Plots Unlock Time-Series Understanding in Multimodal Modelscited by 9
  9. 2025TS-Agent: Understanding and Reasoning Over Raw Time Series via Iterative Insight Gatheringcited by 9
  10. 2026MMTS-BENCH: A Comprehensive Benchmark for Time Series Understanding and Reasoningcited by 9
  11. 2026ARFBench: Benchmarking Time Series Question Answering Ability for Software Incident Responsecited by 6
  12. 2026TimeSage-MT: A Multi-Turn Benchmark for Evaluating Agentic Time Series Reasoningcited by 5
  13. 2024Measuring Agreeableness Bias in Multimodal Modelscited by 4
  14. 2026Seeking SOTA: Time-Series Forecasting Must Adopt Taxonomy-Specific Evaluation to Dispel Illusory Gainscited by 3
  15. 2026Are Tools Always Beneficial? Learning to Invoke Tools Adaptively for Dual-Mode Multimodal LLM Reasoningcited by 2
  16. 2026Representing Time Series as Structured Programs for LLM Reasoningcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026TS-Skill: A Benchmark for Evaluating Analytical Skills in Time-Series Question Answering
  2. 2026Thoth: Mid-Training Bridges LLMs to Time Series Understanding
  3. 2025Chat-TS: Enhancing Multi-Modal Reasoning Over Time-Series and Natural Language Data
  4. 2026Zero-Shot Time-Series Question Answering via Decoupled Perception and Reasoning
  5. 2026Lapras: Latent Reasoning for Time Series Language Models
  6. 2026From Recognition to Understanding: Unlocking Cognitive Time Series Reasoning with LLMs
  7. 2026TimeThink: Eliciting Compositional Reasoning in Timeseries Large Language Models
  8. 2024Towards Time Series Reasoning with LLMs