ttda704 at SemEval-2026 Task 6: 用于政治回避检测的结构化思维链提示
ttda704 at SemEval-2026 Task 6: Structured Chain-of-Thought Prompting for Political Evasion Detection
AI总结：
针对总统访谈中政治回避策略分类任务，比较QLoRA微调Qwen3与结构化思维链提示DeepSeek-V3.2/Grok-4-Fast，发现后者在Macro F1上显著更优，最佳系统在9类回避任务上Macro F1达0.5147。
摘要：
本文描述了我们在SemEval-2026任务6中的系统，该任务涉及对从美国总统访谈中提取的英文问答对进行政治回避策略分类。我们系统比较了两种不同的范式：(1) 使用QLoRA对Qwen3模型（4B-32B）进行参数高效微调，通过分层上采样和加权交叉熵损失来应对严重的类别不平衡；(2) 对具备推理能力的API模型（即DeepSeek-V3.2和Grok-4-Fast）使用结构化思维链（CoT）提示。我们的评估表明，启用推理能力的模型的结构化CoT提示在绝对Macro F1上显著优于我们的基线参数高效微调实现。我们最好的系统，即具有扩展推理和少样本分层CoT提示的Grok-4-Fast，在子任务2（9类回避）上达到0.5147的Macro F1，在子任务1（3类清晰度）上达到0.7979的Macro F1，在官方排行榜上分别位列子任务2的第8名（共33支队伍）和子任务1的第13名（共41支队伍）。此外，我们的消融研究揭示了回避检测中有效提示设计的关键见解：在分层分类法中呈现标签有助于结构化模型推理，而少样本示例提供了任务校准。然而，最强的提示变体在Macro F1上并无统计显著差异，而显式启用扩展推理模式通过促进检测回避意图所需的多步语用分析，带来了显著的性能提升。
英文摘要：
This paper describes our system for SemEval-2026 Task 6, which addresses the classification of political evasion strategies in English question-answer pairs extracted from U.S. presidential interviews. We systematically compare two distinct paradigms: (1) Parameter-Efficient Fine-Tuning of Qwen3 models (4B-32B) using QLoRA, enhanced with tiered upsampling and weighted cross-entropy loss to address severe class imbalance, and (2) structured Chain-of-Thought (CoT) prompting of reasoning-capable API models, namely DeepSeek-V3.2 and Grok-4-Fast. Our evaluation demonstrates that structured CoT prompting of reasoning-enabled models substantially outperforms our baseline parameter-efficient fine-tuning implementation in absolute Macro F1. Our best system, Grok-4-Fast with extended reasoning and few-shot hierarchical CoT prompting, achieves a Macro F1 of 0.5147 on Subtask 2 (9-class evasion) and 0.7979 on Subtask 1 (3-class clarity), ranking 8th out of 33 teams on Subtask 2 and 13th out of 41 teams on Subtask 1 on the official leaderboard. Furthermore, our ablation studies reveal key insights into effective prompt design for evasion detection: presenting labels within a hierarchical taxonomy helps structure model reasoning, while few-shot exemplars provide task calibration. However, the strongest prompt variants are not statistically distinguishable in Macro F1, and explicitly enabling extended reasoning modes yields substantial performance gains by facilitating the multi-step pragmatic analysis required to detect evasive intent.
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。
