探究优化下上下文与参数化思维链忠实性之间的相互作用
Investigating the Interplay between Contextual and Parametric Chain-of-Thought Faithfulness under Optimization
AI总结：
通过提出统一偏好对齐接口FaithMate，研究上下文与参数化两种思维链忠实性范式在优化下的相互作用，发现两者正相关但不对称，且上下文忠实性指标间存在权衡。
摘要：
思维链（CoT）忠实性，即CoT是否真实反映大型语言模型（LLM）的底层行为，通常通过两种不相交的范式进行评估：上下文忠实性（通过扰动输入或CoT轨迹测量）和参数化忠实性（通过干预模型的参数化知识评估）。然而，先前的工作仅对它们进行描述性比较。我们通过提出FaithMate（一个统一的偏好对齐接口，用于优化模型朝向任一忠实性范式）来填补这一空白。它使我们能够研究两种范式之间的相互作用，检查忠实性增益在范式内部和跨范式之间是否以及多大程度上泛化。在三个模型、两个数据集和六个忠实性指标上，我们发现两种范式呈正相关但不对称：优化参数化忠实性在两种范式上均产生一致的增益，而上下文对应范式则带来更多可变的增益。在上下文范式内，一个指标上的忠实性增益不能一致地转移到其他指标上，这表明现有的上下文指标捕捉了忠实性的不同方面，并暴露了固有的权衡。这些发现意味着CoT忠实性不是一个单一目标，因此需要多方面的优化和评估。
英文摘要：
Chain-of-Thought (CoT) faithfulness, i.e., whether CoTs genuinely reflect large language models' (LLM) underlying behavior, is typically evaluated with metrics under two disjoint paradigms: contextual faithfulness, measured by perturbing the input or CoT trace, and parametric faithfulness, assessed by intervening on a model's parametric knowledge. Yet prior work compares them only descriptively. We fill this gap by proposing FaithMATE, a unified preference-alignment interface for optimizing models towards either faithfulness paradigm. It enables us to investigate the interplay between the two paradigms, examining whether and to what extent faithfulness gains generalize within and across paradigms. Across three models, two datasets, and six faithfulness metrics, we find that the two paradigms are positively coupled, yet asymmetric: optimizing towards parametric faithfulness yields consistent gains across both paradigms, whereas the contextual counterpart delivers more variable gains. Within the contextual paradigm, faithfulness gains on one metric do not consistently transfer to others, implying that existing contextual metrics capture disjoint facets of faithfulness and exposing inherent trade-offs. These findings imply that CoT faithfulness is not a monolithic objective and therefore requires multifaceted optimization and evaluation.
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。
