Research
How Robust Is Multimodal Claim Verification to LLM Rewriting?
arXiv:2610.02841v1 Announce Type: new Abstract: LLMs are known to introduce stylistic changes into generated text, yet how these stylistic shifts affect model decisions on scientific tasks remains und
arXiv:2610.02841v1 Announce Type: new Abstract: LLMs are known to introduce stylistic changes into generated text, yet how these stylistic shifts affect model decisions on scientific tasks remains underexplored. In this paper, we focus on multimodal claim verification, where the goal is to determine whether a textual claim is grounded in a given piece of evidence. We apply two rewriting strategies: natural rewriting, which simulates how researchers routinely use LLMs to polish academic text, and controlled injection, which inserts a single LLM-associated word to isolate the effect of vocabulary choice. We evaluate 11 open-weight models spanning five VLM families and ranging from 2B to 38B parameters. We find that models are robust to these modifications: most show no significant drop in accuracy, and compared to prior work on review-score manipulation, verification appears far more stable. However, consistent probability shifts do occur. Hedging-oriented conditions produce significant shifts across nearly all models, while boosting conditions show a weaker effect and general polishing conditions (e.g., grammar correction, fluency improvement) have little effect.
Related
- Encoded but Not Routed: Explaining the Table-Chart Gap in Scientific Claim Verification
- From Tables to Quantified Statements: Evaluating LLM Inference Generation through Executable Verification
- LaTER: Efficient Test-Time Reasoning via Latent Exploration and Explicit Verification
Source: arXiv cs.CL | 2026-10-05