Papers
arxiv:2510.13940

Less is More: Improving LLM Reasoning with Minimal Test-Time Intervention

Published on Oct 15
· Submitted by Zhen Yang on Oct 16
Authors:
,
,
,
,
,
,

Abstract

Minimal Test-Time Intervention (MTI) enhances reasoning accuracy and stability in large language models with minimal overhead by selectively applying classifier-free guidance and lightweight negative-prompt guidance.

AI-generated summary

Recent progress in large language models (LLMs) has focused on test-time scaling to improve reasoning via increased inference computation, but often at the cost of efficiency. We revisit test-time behavior and uncover a simple yet underexplored phenomenon: reasoning uncertainty is highly localized-only a small subset of high-entropy tokens dominantly affects output correctness. Motivated by this, we propose Minimal Test-Time Intervention (MTI), a training-free framework that enhances reasoning accuracy and stability with minimal overhead. MTI includes: (i) Selective CFG intervention, applying classifier-free guidance only at uncertain positions; and (ii) Lightweight negative-prompt guidance, reusing the main model's KV cache to approximate unconditional decoding efficiently. MTI yields consistent gains across general, coding, and STEM tasks-e.g., +1.35% average improvement on eight benchmarks for Qwen3-8B-Base and +5% on AIME2024 using Qwen3-32B-Reasoning-while remaining highly efficient.

Community

Paper author Paper submitter

The code is open source; the repository is at https://github.com/EnVision-Research/MTI.

Sign up or log in to comment

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2510.13940 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2510.13940 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2510.13940 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.