Externalized CPDAG Summaries Improve LLM Causal Deduction

arXiv:2609.31071v1 Announce Type: new Abstract: Corr2Cause asks whether a causal claim holds in every DAG compatible with observed correlations and conditional independencies. We frame this as latent-object reasoning: the label is defined by a CPDAG query, but free-form chain-of-thought often collapses the Markov-equivalence-class problem into local pattern matching. We propose Structured Thinking, a two-turn pipeline that first externalizes a typed, schema-constrained CPDAG summary and then answers against that graph state. On the Corr2Cause full test, Structured Thinking raises Qwen3.5-27B from 73.0 to 86.4 F_1(Yes) over a strong PC-instruction baseline in the primary paired run (+13.4 pp; McNemar p=2.4imes 10^{-6}; bootstrap 95% CI [+8.4, +18.6]); across three full-ID seeds, the mean gain is +8.1 pm 5.3 pp. A PC-scaffolded two-turn prose control reaches only 67.6 F_1, indicating that a detailed PC scaffold plus a schema-free prose intermediate is not sufficient. The same pattern holds on Qwen3.6-27B, Paraphrase-OOD, and GPT-5.4-mini. Scrambling the emitted CPDAG costs 12.0 pp F_1, and a full-split audit shows close agreement with the reference CPDAG (ID skeleton F_1 0.960; exact match 75.9%). These results support a bounded design principle: externalize the latent object that defines the label, constrain its form, and test whether downstream answers use it.

Related

• When and Why LLM Causal Priors Help: Closed-Loop Prior Selection for Amortized Causal Inference
• Does Topic Sentiment Cause Perceived Ideology? Comparing Human and LLM Annotations in Political News Articles
• Fair outputs, Biased Internals: Causal Potency and Asymmetry of Latent Bias in LLMs for High-Stakes Decisions

Source: arXiv cs.AI | 2026-09-28