Year: 2026
Type: conference-paper
Abstract: Large Language Models (LLMs) achieve strong reasoning performance, but their robustness to realistic lexical corruption remains poorly understood.We evaluate four openweight instruction-tuned models and frontier models across four reasoning benchmarks under keyboard noise, character swaps, and filler insertion.Character-level perturbations substantially degrade accuracy, especially on multistep reasoning tasks, while filler insertion has little effect.We trace this asymmetry to Attention Diversion: lexical corruption fragments subword tokenization, and the resulting fragments attract disproportionate attention mass, concentrated in middle and final transformer layers.Length-matched controls confirm that fragmentation, not prompt length, drives the loss.A factorial intervention then shows why the damage is hard to undo: fragmentation corrupts token content and attention allocation together, and the two are coupled.Restoring clean attention while the content remains corrupted is actively harmful, restoring content alone is insufficient, and only restoring both recovers a substantial share of the gap.This coupling explains why inference-time strategies, including chain-of-thought prompting, spellchecking, self-repair, and stronger repair models, fail to consistently recover performance: each addresses one channel at a time.
Authors: Jiaqian Zhu, Yang Zhang, Junhua Ding, Xiaowei Yu
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7204182662