Adaptive Rectification Sampling for Test-Time Compute Scaling
This work addresses inefficiencies in test-time compute scaling for LLMs, offering a domain-specific improvement for logical reasoning tasks.
The paper tackles the problem of token waste and reduced readability in test-time scaling methods for large language models by proposing Adaptive Rectification Sampling (AR-Sampling), which guides models to self-correct at appropriate steps, improving accuracy on GSM8K and MATH500 benchmarks while generating a reasonable number of additional tokens.
The newly released OpenAI-o1 and DeepSeek-R1 have demonstrated that test-time scaling can significantly improve model performance, especially in complex tasks such as logical reasoning. Common test-time scaling methods involve generating more chains of thought (CoTs) or longer CoTs with self-correction. However, while self-correction can improve performance, it may lead to significant token waste and reduce readability of the CoT if the reasoning steps are already correct. To demonstrate that large language models (LLMs) can rectify errors at a more fine-grained level, we propose Adaptive Rectification Sampling (AR-Sampling), which can guide the LLMs to self-correction at the appropriate step. AR-Sampling leverages a process-supervised reward model (PRM) as a verifier and constructed trigger sentences to guide the model in adaptive step-level rethinking. Through the experiments on GSM8K and MATH500, it indicates that our approach enables the models to rethink in more fine-grained level, improving the accuracy of solutions, while generating a reasonable number of additional tokens.