发言前打分：使用回复质量分数提升对话生成中的人设一致性
Score Before You Speak: Improving Persona Consistency in Dialogue Generation using Response Quality Scores
AI总结：
针对大型语言模型对话中人设一致性难的问题，提出SBS框架，在训练阶段统一回复与质量分数的学习，并在推理时利用该知识，实验证明其有效提升了百万至十亿参数模型的人设一致性。
摘要：
基于人设的对话生成是构建对话人工智能的重要里程碑。尽管大型语言模型（LLMs）的能力不断提升，但由于现有对话数据的多样性有限，在对话中有效融入人设保真度仍具挑战性。我们提出了一种新颖的框架 SBS（Score-Before-Speaking），它优于先前的方法，并为百万和十亿参数模型均带来了提升。与先前的方法不同，SBS 将回复及其相对质量的学习统一到单一步骤中。关键创新在于训练对话模型在训练期间将增强的回复与质量分数相关联，并在推理时利用该知识。我们使用基于名词的替换进行增强，并使用基于语义相似度的分数作为回复质量的代理。通过使用基准数据集（PERSONA-CHAT 和 ConvAI2）进行的大量实验，我们表明条件化打分训练使现有模型能够更好地捕捉一系列人设一致的对话。我们的消融研究还证明，在训练期间将分数包含在输入提示中优于传统训练设置。代码及更多细节请见 https://arpita2512.github.io/score_before_you_speak
英文摘要：
Persona-based dialogue generation is an important milestone towards building conversational artificial intelligence. Despite the ever-improving capabilities of large language models (LLMs), effectively integrating persona fidelity in conversations remains challenging due to the limited diversity in existing dialogue data. We propose a novel framework SBS (Score-Before-Speaking), which outperforms previous methods and yields improvements for both million and billion-parameter models. Unlike previous methods, SBS unifies the learning of responses and their relative quality into a single step. The key innovation is to train a dialogue model to correlate augmented responses with a quality score during training and then leverage this knowledge at inference. We use noun-based substitution for augmentation and semantic similarity-based scores as a proxy for response quality. Through extensive experiments with benchmark datasets (PERSONA-CHAT and ConvAI2), we show that score-conditioned training allows existing models to better capture a spectrum of persona-consistent dialogues. Our ablation studies also demonstrate that including scores in the input prompt during training is superior to conventional training setups. Code and further details are available at https://arpita2512.github.io/score_before_you_speak
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。
