TrajSelector: Harnessing Latent Representations for Efficient and Effective Best-of-N in Large Reasoning Model

About this episode

🤗 Upvotes: 32 | cs.CL

        <strong>Authors:</strong><br />
        Bin Yu, Xinming Wang, Shijie Lian, Haotian Li, Changti Wu, Ruina Hu, Bailing Wang, Yuliang Wei, Kai Chen</p>

        <strong>Title:</strong><br />
        TrajSelector: Harnessing Latent Representations for Efficient and Effective Best-of-N in Large Reasoning Model</p>

        <strong>Arxiv:</strong><br />
        <a href="http://arxiv.org/abs/2510.16449v1">http://arxiv.org/abs/2510.16449v1</a></p>

        <strong>Abstract:</strong><br />
        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