发表机构
Fondazione Bruno Kessler; University of Groningen（布鲁诺·凯塞勒基金会; 格罗宁根大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究通过分离训练、推理、评估三阶段的对话者传记可见性，发现训练阶段可见性对基于角色的对话生成影响更大，揭示了传记泄露的成因。
AI 中文摘要
基于角色的对话系统通常以说话人传记为条件，但对话至少涉及两名参与者，谁能访问谁的传记会因训练、推理和评估阶段而异。现有研究常忽略这些方面，模糊了仅在训练、推理和评估阶段分别切换传记可见性时才会显现的机制，而这种三阶段分解在现有研究中大多被视为单一因素。我们在一个配有说话人传记的对话数据集上研究该分解，改变训练和推理阶段中目标说话人与对话者是否能看到彼此的传记，并使用大语言模型（LLM）作为评判者执行作者识别任务。我们发现：（i）训练阶段的可见性比推理阶段更能决定模型是通过对话表达角色特质，还是退而复制传记文本（这是基于角色生成中已知的问题/现象）；（ii）使用对话者-传记可见性训练的模型，比未使用该可见性训练的模型复制的目标传记文本更少，而仅在推理阶段改变可见性的效果一致性较差；（iii）在非对称披露场景下，即仅对话者能看到目标传记时，目标内容更常泄露到对话者的轮次中，且包含此类痕迹的对话更易被评判者识别，尤其是当对话者的轮次可见时。这些结果表明，生成轮次中的传记泄露是训练和推理阶段对话者可见性配置方式的产物，因此有必要对这三个阶段进行分离。
英文摘要
Persona-based dialogue systems are usually conditioned on speaker biography, but dialogues involve at least two participants, and who has access to whose biography can vary across training, inference, and evaluation. Prior work often neglected these aspects, obscuring mechanisms that only appear when biography visibility is toggled separately across training, inference, and evaluation, a three-stage factorisation that prior work has largely treated as a single factor. We study this factorisation on a dataset of dialogues paired with speaker's biographies, varying whether the target and interlocutor speakers see each other's biographies during training and inference, and using an LLM as a judge to perform author identification. We find that (i) training-time visibility, more than inference-time visibility, determines whether models express persona traits through dialogue or fall back on copying biographical text (a known problem/phenomenon in persona-based generation); (ii) models trained with interlocutor-biography visibility copy less target-biographical text than models trained without it, while changing visibility only at inference time has a less consistent effect; and (iii) under asymmetric disclosure, where only the interlocutor sees the target biography, target content leaks into interlocutor turns more often, and dialogues containing such traces are easier for the judge to identify, especially when interlocutor turns are visible. These results suggest that biography leakage into generated turns is an artefact of how interlocutor visibility is configured across training and inference, and separating the three stages is necessary.
