English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Prompt-based learning has been an effective paradigm for large pretrainedlanguage models (LLM), enabling few-shot or even zero-shot learning. Black-boxprompt search has received growing interest recently for its distinctiveproperties of gradient-free optimization, proven particularly useful andpowerful for model-as-a-service usage. However, the discrete nature and thecomplexity of combinatorial optimization hinder the efficiency of modernblack-box approaches. Despite extensive research on search algorithms, thecrucial aspect of search space design and optimization has been largelyoverlooked. In this paper, we first conduct a sensitivity analysis by promptingLLM, revealing that only a small number of tokens exert a disproportionateamount of influence on LLM predictions. Leveraging this insight, we propose theClustering and Pruning for Efficient Black-box Prompt Search (ClaPS), a simpleblack-box search method that first clusters and prunes the search space tofocus exclusively on influential prompt tokens. By employing even simple searchmethods within the pruned search space, ClaPS achieves state-of-the-artperformance across various tasks and LLMs, surpassing the performance ofcomplex approaches while significantly reducing search costs. Our findingsunderscore the critical role of search space design and optimization inenhancing both the usefulness and the efficiency of black-box prompt-basedlearning.

Survival of the Most Influential Prompts: Efficient Black-Box Prompt Search via Clustering and Pruning