用图检索+反思式提示，让大模型更准地推理事件因果。
AILS-NTUA at SemEval-2026 Task 12: Graph-Based Retrieval and Reflective Prompting for Abductive Event Reasoning
- 三阶段流程：图检索找线索，大模型推理，后处理校正一致性。
- 准确率0.95，击败14个模型，跨模型错误模式高度一致。
- 揭示因果推理共性缺陷，适合研究大模型推理偏差的学者。
我们提出一种在SemEval-2026任务12中夺冠的三阶段系统，用于归纳性事件推理：结合基于图的检索、通过反思式提示优化的大模型推理，以及事后一致性强化。该系统在评估阶段排行榜上排名第一，准确率达0.95。对14个模型（7个模型家族）的跨模型错误分析揭示了三种共有的归纳偏差：因果链不完整、偏好近因、显著性偏差；这些偏差在多标签因果推理中的跨家族收敛（原因数量减少51%），表明存在系统性而非模型特异性的失败模式。
原文摘要 · Abstract (English)
We present a winning three-stage system for SemEval 2026 Task~12: Abductive Event Reasoning that combines graph-based retrieval, LLM-driven abductive reasoning with prompt design optimized through reflective prompt evolution, and post-hoc consistency enforcement; our system ranks first on the evaluation-phase leaderboard with an accuracy score of 0.95. Cross-model error analysis across 14 models (7~families) reveals three shared inductive biases: causal chain incompleteness, proximate cause preference, and salience bias, whose cross-family convergence (51\% cause-count reduction) indicates systematic rather than model-specific failure modes in multi-label causal reasoning.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。