TrajSelector: Harnessing Latent Representations for Efficient and Effective Best-of-N in Large Reasoning Model
This addresses efficiency and performance issues in large reasoning models for AI researchers and practitioners, offering an incremental improvement over existing Best-of-N frameworks.
The paper tackles the computational overhead and underutilization of latent representations in Best-of-N selection for large language models by introducing TrajSelector, which uses a lightweight verifier to score reasoning trajectories based on hidden states, achieving accuracy gains of 4.31% to 12.21% over existing methods while reducing inference costs.
Large language models (LLMs) have shown remarkable progress in complex reasoning tasks, largely enabled by test-time scaling (TTS) paradigms that allocate additional compute during inference. Among these, external TTS (particularly the Best-of-N selection paradigm) yields scalable performance improvements by selecting from multiple independently generated reasoning trajectories. However, this approach faces key limitations: (i) the high computational overhead of deploying process reward models, (ii) the underutilization of the LLM's intrinsic latent representations. We introduce TrajSelector, an efficient and effective Best-of-N framework that exploit the hidden states in the sampler LLM for process-level scoring. A lightweight verifier (with only 0.6B parameters) evaluates the quality of step-wise trajectory, and then aggregates these scores to identify the optimal reasoning trajectory. Our framework employs a fully data-driven, end-to-end training recipe that eliminates reliance on massive step-level annotations. Experiential results across five benchmarks demonstrate that TrajSelector delivers consistent performance gains. In Best-of-32 settings, it surpasses majority voting by 4.61% accuracy and outperforms existing process reward models by 4.31% to 12.21%, all while maintaining lower inference costs.