Pith. sign in

Reasoning Robustness of LLMs to Adversarial Typographi- cal Errors

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

citation-role summary

background 1

citation-polarity summary

fields

cs.CR 1

years

2025 1

verdicts

CONDITIONAL 1

roles

background 1

polarities

support 1

representative citing papers

Excessive Reasoning Attack on Reasoning LLMs

cs.CR · 2025-06-17 · conditional · novelty 6.0

A 10-token adversarial suffix, optimized with three custom losses, raises reasoning length 3x to 9x on open reasoning LLMs without hurting accuracy, and partly transfers to closed models.

citing papers explorer

Showing 1 of 1 citing paper.

  • Excessive Reasoning Attack on Reasoning LLMs cs.CR · 2025-06-17 · conditional · none · ref 9

    A 10-token adversarial suffix, optimized with three custom losses, raises reasoning length 3x to 9x on open reasoning LLMs without hurting accuracy, and partly transfers to closed models.