Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2602.06373 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911427155984384 |
|---|---|
| author | Yan, Tianqiang Shang, Sihan Li, Yuheng Qiu, Song Peng, Hao Luo, Wenjian Xie, Jue Qu, Lizhen Gao, Yuan |
| author_facet | Yan, Tianqiang Shang, Sihan Li, Yuheng Qiu, Song Peng, Hao Luo, Wenjian Xie, Jue Qu, Lizhen Gao, Yuan |
| contents | While self-reflection can enhance language model reliability, its underlying mechanisms remain opaque, with existing analyses often yielding correlation-based insights that fail to generalize. To address this, we introduce \textbf{\texttt{ReBeCA}} (self-\textbf{\texttt{Re}}flection \textbf{\texttt{Be}}havior explained through \textbf{\texttt{C}}ausal \textbf{\texttt{A}}nalysis), a framework that unveils the interpretable behavioral hierarchy governing the self-reflection outcome. By modeling self-reflection trajectories as causal graphs, ReBeCA isolates genuine determinants of performance through a three-stage Invariant Causal Prediction (ICP) pipeline. We establish three critical findings: (1) \textbf{Behavioral hierarchy:} Semantic behaviors of the model influence final self-reflection results hierarchically: directly or indirectly; (2) \textbf{Causation matters:} Generalizability in self-reflection effects is limited to just a few semantic behaviors; (3) \textbf{More $\mathbf{\neq}$ better:} The confluence of seemingly positive semantic behaviors, even among direct causal factors, can impair the efficacy of self-reflection. ICP-based verification identifies sparse causal parents achieving up to $49.6\%$ structural likelihood gains, stable across tasks where correlation-based patterns fail. Intervention studies on novel datasets confirm these causal relationships hold out-of-distribution ($p = .013, η^2_\mathrm{p} = .071$). ReBeCA thus provides a rigorous methodology for disentangling genuine causal mechanisms from spurious associations in self-reflection dynamics. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_06373 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ReBeCA: Unveiling Interpretable Behavior Hierarchy behind the Iterative Self-Reflection of Language Models with Causal Analysis Yan, Tianqiang Shang, Sihan Li, Yuheng Qiu, Song Peng, Hao Luo, Wenjian Xie, Jue Qu, Lizhen Gao, Yuan Computation and Language While self-reflection can enhance language model reliability, its underlying mechanisms remain opaque, with existing analyses often yielding correlation-based insights that fail to generalize. To address this, we introduce \textbf{\texttt{ReBeCA}} (self-\textbf{\texttt{Re}}flection \textbf{\texttt{Be}}havior explained through \textbf{\texttt{C}}ausal \textbf{\texttt{A}}nalysis), a framework that unveils the interpretable behavioral hierarchy governing the self-reflection outcome. By modeling self-reflection trajectories as causal graphs, ReBeCA isolates genuine determinants of performance through a three-stage Invariant Causal Prediction (ICP) pipeline. We establish three critical findings: (1) \textbf{Behavioral hierarchy:} Semantic behaviors of the model influence final self-reflection results hierarchically: directly or indirectly; (2) \textbf{Causation matters:} Generalizability in self-reflection effects is limited to just a few semantic behaviors; (3) \textbf{More $\mathbf{\neq}$ better:} The confluence of seemingly positive semantic behaviors, even among direct causal factors, can impair the efficacy of self-reflection. ICP-based verification identifies sparse causal parents achieving up to $49.6\%$ structural likelihood gains, stable across tasks where correlation-based patterns fail. Intervention studies on novel datasets confirm these causal relationships hold out-of-distribution ($p = .013, η^2_\mathrm{p} = .071$). ReBeCA thus provides a rigorous methodology for disentangling genuine causal mechanisms from spurious associations in self-reflection dynamics. |
| title | ReBeCA: Unveiling Interpretable Behavior Hierarchy behind the Iterative Self-Reflection of Language Models with Causal Analysis |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2602.06373 |