用教师示范提升婴儿语言模型的对话连续性
Teacher Demonstrations in a BabyLM's Zone of Proximal Development for Contingent Multi-Turn Interaction
- 设计教师-学生框架，针对婴儿语言模型优化多轮对话连续性
- 在1亿词语料上训练的模型对话更语法正确、连贯
- 适合研究儿童语言发展与对话系统对齐的学者
儿童与照料者之间的多轮对话具有连续性特征——即对话双方能及时、直接且有意义地互动。本文提出ContingentChat框架，用于评估并改进在1亿词语料上训练的BabyLM的多轮对话连续性。通过引入新颖的后训练对齐数据集，使BabyLM生成的回应更具语法正确性和连贯性。实验表明，采用自适应教师解码策略仅带来有限提升。结果表明，针对性后训练能有效改善对话质量，但连续性仍是BabyLM面临的重要挑战。
原文摘要 · Abstract (English)
Multi-turn dialogues between a child and a caregiver are characterized by a property called contingency - that is, prompt, direct, and meaningful exchanges between interlocutors. We introduce ContingentChat, a teacher-student framework that benchmarks and improves multi-turn contingency in a BabyLM trained on 100M words. Using a novel alignment dataset for post-training, BabyLM generates responses that are more grammatical and cohesive. Experiments with adaptive teacher decoding strategies show limited additional gains. ContingentChat demonstrates the benefits of targeted post-training for dialogue quality and indicates that contingency remains a challenging goal for BabyLMs.
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。