Sticker-TTS: Learn to Utilize Historical Experience with a Sticker-driven Test-Time Scaling Framework

ACL ARR 2025 May Submission7725 Authors

Abstract: Large reasoning models (LRMs) have exhibited strong performance on complex reasoning tasks, with further gains achievable through increased computational budgets at inference. However, current test-time scaling methods predominantly rely on redundant sampling, ignoring the historical experience utilization, thereby limiting computational efficiency. To overcome this limitation, we propose \textbf{Sticker-TTS}, a novel test-time scaling framework that coordinates three collaborative LRMs to iteratively explore and refine solutions guided by historical attempts. At the core of our framework are distilled key conditions—termed \textit{stickers}—which drive the extraction, refinement, and reuse of critical information across multiple rounds of reasoning. To further enhance the efficiency and performance of our framework, we introduce a two-stage optimization strategy that combines imitation learning with self-improvement, enabling progressive refinement. Extensive evaluations on three challenging mathematical reasoning benchmarks, including AIME-24, AIME-25, and OlymMATH, demonstrate that Sticker-TTS consistently surpasses strong baselines, including self-consistency and advanced reinforcement learning approaches, under comparable inference budgets. These results highlight the effectiveness of sticker-guided historical experience utilization.

Paper Type: Long

Research Area: NLP Applications

Research Area Keywords: mathematical NLP

Contribution Types: NLP engineering experiment

Languages Studied: English

Submission Number: 7725