在推理过程中增强大语言模型的推理准确性
Enhancing reasoning accuracy in large language models during inference time
AI总结：
本文研究了提升大语言模型推理准确性的推理时技术，通过自一致性、双模型推理一致性和自反思三种策略，发现自一致性在核采样和受控温度下能显著提高准确性，适用于低风险领域。
摘要：
大型语言模型（LLMs）在语言能力方面表现强大，但在多步骤推理任务中往往不可靠，尤其是在未进行额外训练或微调时。本文研究了推理时的技术以提高LLMs的推理准确性。我们系统评估了三种推理时策略：（i）通过随机解码实现的自一致性，其中模型使用受控温度和核采样多次采样，并选择最频繁的最终答案；（ii）双模型推理一致，其中两个独立模型的输出被比较，只有一致的推理轨迹才被信任；（iii）自反思，其中模型批评并修改其自身的推理。在所有评估方法中，我们使用链式思维（CoT）[1]提示来在生成最终答案前引出显式的中间推理步骤。在本文中，我们提供了在相同提示和验证设置下的三种推理策略的受控比较评估。我们的实验表明，使用核采样和受控温度值的自一致性方法能显著提高准确性，比贪心单次解码在准确性上提高9%至15%，适用于低风险领域，提供有意义的提升，且开销小。双模型方法为模型推理步骤提供额外确认，因此更适合中等风险领域，其中更高的可靠性justify额外的计算。自反思仅提供有限的改进，表明在推理时对较小的非推理模型效果有限。
英文摘要：
Large Language Models (LLMs) often exhibit strong linguistic abilities while remaining unreliable on multi-step reasoning tasks, particularly when deployed without additional training or fine-tuning. In this work, we study inference-time techniques to improve the reasoning accuracy of LLMs. We systematically evaluate three classes of inference-time strategies: (i) self-consistency via stochastic decoding, where the model is sampled multiple times using controlled temperature and nucleus sampling and the most frequent final answer is selected; (ii) dual-model reasoning agreement, where outputs from two independent models are compared and only consistent reasoning traces are trusted; and (iii) self-reflection, where the model critiques and revises its own reasoning. Across all evaluated methods, we employ Chain-of-Thought (CoT) [1] prompting to elicit explicit intermediate reasoning steps before generating final answers. In this work, we provide a controlled comparative evaluation across three inference-time strategies under identical prompting and verification settings. Our experiments on LLM [2] show that self-consistency with nucleus sampling and controlled temperature value yields the substantial gains, achieving a 9% to 15% absolute improvement in accuracy over greedy single-pass decoding, well-suited for low-risk domains, offering meaningful gains with minimal overhead. The dual-model approach provides additional confirmation for model reasoning steps thus more appropriate for moderate-risk domains, where higher reliability justifies additional compute. Self-reflection offers only marginal improvements, suggesting limited effectiveness for smaller non-reasoning models at inference time.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
