Can Language Models Take A Hint? Prompting for Controllable Contextualized Commonsense Inference

TL;DR

This paper introduces 'hinting,' a prompt-based data augmentation method that improves controllability in contextual commonsense inference by guiding language models without sacrificing performance.

Contribution

It proposes a novel hinting technique using hard and soft prompts to enhance controllability in commonsense inference tasks.

Findings

Hinting improves control over inference focus.

Performance remains stable with hinting.

Synonyms and antonyms enhance hinting effectiveness.

Abstract

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.