mlsanityData to 14 Oct 2026

DS@GT at CheckThat! 2025: Evaluating Context and Tokenization Strategies for Numerical Fact Verification

arXiv 2507.06195 ↗First posted 8 Jul 2025cs.CLcited by 2

Not on the map. Why?

Abstract (arXiv)

Numerical claims, statements involving quantities, comparisons, and temporal references, pose unique challenges for automated fact-checking systems. In this study, we evaluate modeling strategies for veracity prediction of such claims using the QuanTemp dataset and building our own evidence retrieval pipeline. We investigate three key factors: (1) the impact of more evidences with longer input context windows using ModernBERT, (2) the effect of right-to-left (R2L) tokenization, and (3) their combined influence on classification performance. Contrary to prior findings in arithmetic reasoning tasks, R2L tokenization does not boost natural language inference (NLI) of numerical tasks. A longer context window does also not enhance veracity performance either, highlighting evidence quality as the dominant bottleneck. Our best-performing system achieves competitive macro-average F1 score of 0.57 and places us among the Top-4 submissions in Task 3 of CheckThat! 2025. Our code is available at https://github.com/dsgt-arc/checkthat-2025-numerical.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 12 of its 21 references that are in the map. First posted on arXiv 8 Jul 2025.

Builds on

12 of 21 references.

References · 12

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  3. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,898
  4. 2016MS MARCO: A Human Generated MAchine Reading COmprehension Datasetcited by 1,605
  5. 2019MathQA: Towards Interpretable Math Word Problem Solving with Operation-Based Formalismscited by 662
  6. 2024Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inferencecited by 554
  7. 2023Complex Claim Verification with Evidence Retrieved in the Wildcited by 110
  8. 2022Generating Literal and Implied Subquestions to Fact-check Complex Claimscited by 66
  9. 2023Information Retrieval: Recent Advances and Beyondcited by 49
  10. 2024QuanTemp: A real-world open-domain benchmark for fact-checking numerical claimscited by 23
  11. 2025The CLEF-2025 CheckThat! Lab: Subjectivity, Fact-Checking, Claim Normalization, and Retrievalcited by 14
  12. 2024rerankers: A Lightweight Python Library to Unify Ranking Methodscited by 8

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Symbolic Augmentation Closes a Canonical-Equivalence Blind Spot in Neural Fact-Checkerscited by 0
  2. 2026DS@GT ARC at CheckThat! 2026: LLM-Based Trace Ranking and Grouped Reward Modeling for Multilingual Numerical Claim Verificationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026R2VC: Modular Fact-Checking with Retrieval, Verification, and Confidence Calibration
  2. 2024Tokenization counts: the impact of tokenization on arithmetic in frontier LLMs
  3. 2025Think Right, Not More: Test-Time Scaling for Numerical Claim Verification
  4. 2026What Do Claim Verification Datasets Actually Test? A Reasoning Trace Analysis
  5. 2026Evaluating and Improving Evidence-Grounded Fact-Checking in LLMs via Multi-Round Evidence Ablation
  6. 2025Fact-checking with Generative AI: A Systematic Cross-Topic Examination of LLMs Capacity to Detect Veracity of Political Information
  7. 2023The Perils & Promises of Fact-checking with Large Language Models
  8. 2026Context Shapes LLMs Retrieval-Augmented Fact-Checking Effectiveness