ReBeCA: 通过因果分析揭示语言模型迭代自我反思背后的可解释行为层级
ReBeCA: Causality-Inspired Behavior Analysis for Understanding the Self-Reflection of Language Models
AI总结:
ReBeCA通过因果分析揭示语言模型自我反思中的可解释行为层级,识别因果关系以提升自我反思效果。
摘要:
尽管自我反思可以提高语言模型的可靠性,但其内在机制仍不清楚,现有分析往往只能提供基于相关性的见解,无法推广。为此,我们引入ReBeCA(通过因果分析解释自我反思行为),该框架揭示了支配自我反思结果的可解释行为层级。通过将自我反思轨迹建模为因果图,ReBeCA通过三阶段不变因果预测(ICP)流程隔离性能的真实决定因素。我们得出三个关键发现:(1)行为层级:模型的语义行为通过直接或间接的方式影响最终的自我反思结果;(2)因果重要:自我反思效果的可推广性仅限于少数语义行为;(3)更多≠更好:看似积极的语义行为的融合,即使在直接因果因素中,也可能损害自我反思的有效性。基于ICP的验证发现,稀疏因果父节点在结构似然上可提高高达49.6%,在相关性失败的任务中保持稳定。在新数据集上的干预研究证实这些因果关系在分布外情况下仍然成立(p=0.013,η²_p=0.071)。ReBeCA因此提供了一种严谨的方法,用于区分自我反思动态中的真实因果机制与虚假关联。
英文摘要:
Self-reflection is a basic self-improvement skill of language agents, yet it does not reliably improve their outputs, where the underlying behavioral mechanism remains opaque. We introduce **ReBeCA** (self-**Re**flection **Be**havior explained through **C**ausality-inspired **A**nalysis), a framework for identifying and validating outcome-relevant reflective behaviors from recorded trajectories. ReBeCA uses **reflective codes**, an inductive coding schema, to represent each trajectory as a sequence of reflective behaviors. It learns dependencies among these behaviors and the outcome with causal discovery, selects the behaviors that best predict the outcome of held-out trajectories, and tests them for consistency across data subsets and by prompt-based steering. In a controlled case study of four Qwen3 models under fixed-round Self-Refine on mathematical reasoning and translation, ReBeCA identifies compact, round-specific behavioral sets with superior held-out predictive fit and uses prompt-based steering on unseen problems to reveal how individual and combined behavioral directives affect performance (Holm-adjusted $p = .020$ on Qwen3-4B). ReBeCA offers a reusable framework that turns agent trajectories into interpretable, verifiable behavioral hypotheses without requiring model internals or preselecting the behaviors to test.
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。