Revisiting Test-Time Scaling: A Survey and a Diversity-Aware Method for Efficient Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chung, Ho-Lam, Hsiao, Teng-Yun, Huang, Hsiao-Ying, Cho, Chunerh, Lin, Jian-Ren, Ziwei, Zhang, Chen, Yun-Nung
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910989122797568
author Chung, Ho-Lam
Hsiao, Teng-Yun
Huang, Hsiao-Ying
Cho, Chunerh
Lin, Jian-Ren
Ziwei, Zhang
Chen, Yun-Nung
author_facet Chung, Ho-Lam
Hsiao, Teng-Yun
Huang, Hsiao-Ying
Cho, Chunerh
Lin, Jian-Ren
Ziwei, Zhang
Chen, Yun-Nung
contents Test-Time Scaling (TTS) improves the reasoning performance of Large Language Models (LLMs) by allocating additional compute during inference. We conduct a structured survey of TTS methods and categorize them into sampling-based, search-based, and trajectory optimization strategies. We observe that reasoning-optimized models often produce less diverse outputs, which limits TTS effectiveness. To address this, we propose ADAPT (A Diversity Aware Prefix fine-Tuning), a lightweight method that applies prefix tuning with a diversity-focused data strategy. Experiments on mathematical reasoning tasks show that ADAPT reaches 80% accuracy using eight times less compute than strong baselines. Our findings highlight the essential role of generative diversity in maximizing TTS effectiveness.
format Preprint
id arxiv_https___arxiv_org_abs_2506_04611
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Revisiting Test-Time Scaling: A Survey and a Diversity-Aware Method for Efficient Reasoning
Chung, Ho-Lam
Hsiao, Teng-Yun
Huang, Hsiao-Ying
Cho, Chunerh
Lin, Jian-Ren
Ziwei, Zhang
Chen, Yun-Nung
Computation and Language
Test-Time Scaling (TTS) improves the reasoning performance of Large Language Models (LLMs) by allocating additional compute during inference. We conduct a structured survey of TTS methods and categorize them into sampling-based, search-based, and trajectory optimization strategies. We observe that reasoning-optimized models often produce less diverse outputs, which limits TTS effectiveness. To address this, we propose ADAPT (A Diversity Aware Prefix fine-Tuning), a lightweight method that applies prefix tuning with a diversity-focused data strategy. Experiments on mathematical reasoning tasks show that ADAPT reaches 80% accuracy using eight times less compute than strong baselines. Our findings highlight the essential role of generative diversity in maximizing TTS effectiveness.
title Revisiting Test-Time Scaling: A Survey and a Diversity-Aware Method for Efficient Reasoning
topic Computation and Language
url https://arxiv.org/abs/2506.04611