通过提示词控制大模型生成常识推理，提升可控性。
Can Language Models Take A Hint? Prompting for Controllable Contextualized Commonsense Inference
- 用硬提示和软提示引导推理方向，实现可控生成。
- 在ParaCOMET和GLUCOSE数据集上保持性能不降。
- 支持通过同义词、反义词增强推理多样性，适合需要精准控制的场景。
在给定故事上下文中生成常识性断言仍是现代语言模型的难题。以往方法通过对齐常识推理与故事内容来训练生成模型，但难以控制生成内容的具体焦点。本文提出“提示引导”（hinting）技术，采用前缀提示策略，结合硬提示与软提示，引导推理过程。我们在两个情境化常识推理数据集——ParaCOMET和GLUCOSE上验证该方法，评估其在通用与特定上下文推理中的表现。进一步实验将同义词与反义词引入提示，考察其影响。结果表明，hinting 在不损害原有推理性能的前提下，显著提升了生成内容的可控性。
原文摘要 · Abstract (English)
Generating commonsense assertions within a given story context remains a difficult task for modern language models. Previous research has addressed this problem by aligning commonsense inferences with stories and training language generation models accordingly. One of the challenges is determining which topic or entity in the story should be the focus of an inferred assertion. Prior approaches lack the ability to control specific aspects of the generated assertions. In this work, we introduce "hinting," a data augmentation technique that enhances contextualized commonsense inference. "Hinting" employs a prefix prompting strategy using both hard and soft prompts to guide the inference process. To demonstrate its effectiveness, we apply "hinting" to two contextual commonsense inference datasets: ParaCOMET and GLUCOSE, evaluating its impact on both general and context-specific inference. Furthermore, we evaluate "hinting" by incorporating synonyms and antonyms into the hints. Our results show that "hinting" does not compromise the performance of contextual commonsense inference while offering improved controllability.
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。