Paraphrase Types Elicit Prompt Engineering Capabilities
This work addresses the problem of prompt sensitivity for users of language models, offering incremental insights into optimizing prompts through linguistic adjustments.
The study systematically evaluated how linguistic variations in prompts affect language models, finding that specific paraphrase types like morphology and lexicon can improve task performance, with median gains of 6.7% in Mixtral 8x7B and 5.5% in LLaMA 3 8B.
Much of the success of modern language models depends on finding a suitable prompt to instruct the model. Until now, it has been largely unknown how variations in the linguistic expression of prompts affect these models. This study systematically and empirically evaluates which linguistic features influence models through paraphrase types, i.e., different linguistic changes at particular positions. We measure behavioral changes for five models across 120 tasks and six families of paraphrases (i.e., morphology, syntax, lexicon, lexico-syntax, discourse, and others). We also control for other prompt engineering factors (e.g., prompt length, lexical diversity, and proximity to training data). Our results show a potential for language models to improve tasks when their prompts are adapted in specific paraphrase types (e.g., 6.7% median gain in Mixtral 8x7B; 5.5% in LLaMA 3 8B). In particular, changes in morphology and lexicon, i.e., the vocabulary used, showed promise in improving prompts. These findings contribute to developing more robust language models capable of handling variability in linguistic expression.