多轮共情对话中的语篇多样性
Discourse Diversity in Multi-Turn Empathic Dialogue
AI总结：
研究探讨了大型语言模型在多轮共情对话中语篇策略的多样性问题，提出MINT框架通过强化学习提升对话多样性，显著提高共情效果并减少重复。
摘要：
大型语言模型（LLMs）在单轮对话中能产生被评价为高度共情的回应，但它们也以重复相同的词汇模式、句法模板和语篇结构著称（Jiang et al., 2025; Shaib et al., 2024; Namuduri et al., 2025）。较少关注的是这种公式化是否延伸到语篇策略层面，即回应对被回应者所做的事情。这对共情对话尤其重要，因为有效的支持不仅需要在某一刻有恰当的回应，还需要在对话展开时采用多样的策略（Stiles et al., 1998）。确实，先前研究表明，在单轮设置中，LLMs重用的战术序列比人类支持者更多（Gueorguieva et al., 2026）。我们扩展了这种分析到多轮对话中，并发现这种僵化现象加剧：一旦一个战术出现在支持者回合中，LLMs在下一个回合中几乎以人类的两倍速度重复使用它（0.50-0.56 vs. 0.27）。这种模式在LLMs作为支持者在真实情感支持对话中也成立，并且无法被标准相似性度量检测到。为解决这一差距，我们引入MINT（多轮跨战术新颖性训练），这是首个优化多轮共情对话语篇策略多样性的强化学习框架。最佳的MINT变体结合了共情质量奖励和跨回合战术新颖性信号，在170亿和400亿参数模型上将整体共情效果提高了25.3%，同时在400亿参数模型上将跨回合语篇策略重复率降低了26.3%，在两项指标上均超越了所有基线方法，包括仅质量或仅词级多样性方法。这些结果表明，当前模型所缺乏的不是共情本身，而是跨对话变化其语篇策略的能力。
英文摘要：
Large language models (LLMs) produce responses rated as highly empathic in single-turn settings (Ayers et al., 2023; Lee et al., 2024), yet they are also known to be formulaic generators that reuse the same lexical patterns, syntactic templates, and discourse structures across tasks (Jiang et al., 2025; Shaib et al., 2024; Namuduri et al., 2025). Less attention has been paid to whether this formulaicity extends to the level of discourse moves, i.e., what a response does for the person it is addressing. This question is especially consequential for empathic dialogue, where effective support demands not just a kind response at one moment but varied strategies as a conversation unfolds (Stiles et al., 1998). Indeed, prior work shows that LLMs reuse the same tactic sequences more than human supporters in single-turn settings (Gueorguieva et al., 2026). We extend this analysis to multi-turn conversations and find that the rigidity compounds: once a tactic appears in a supporter turn, LLMs reuse it in the next at nearly double the rate of humans (0.50-0.56 vs. 0.27). This pattern holds across LLMs serving as supporters in real emotional support conversations, and is invisible to standard similarity metrics. To address this gap, we introduce MINT (Multi-turn Inter-tactic Novelty Training), the first reinforcement learning framework to optimize discourse move diversity across multi-turn empathic dialogue. The best MINT variant combines an empathy quality reward with a cross-turn tactic novelty signal, improving aggregate empathy by 25.3% over vanilla across 1.7B and 4B models while reducing cross-turn discourse move repetition by 26.3% on the 4B model, surpassing all baselines including quality-only and token-level diversity methods on both measures. These results suggest that what current models lack is not empathy itself, but the ability to vary their discourse moves across a conversation.
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。
