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/

Zendy Plus

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

Zendy Tools

Zendy Open

Large Language Models (LLMs) excel in various tasks, but they rely oncarefully crafted prompts that often demand substantial human effort. Toautomate this process, in this paper, we propose a novel framework for discreteprompt optimization, called EvoPrompt, which borrows the idea of evolutionaryalgorithms (EAs) as they exhibit good performance and fast convergence. Toenable EAs to work on discrete prompts, which are natural language expressionsthat need to be coherent and human-readable, we connect LLMs with EAs. Thisapproach allows us to simultaneously leverage the powerful language processingcapabilities of LLMs and the efficient optimization performance of EAs.Specifically, abstaining from any gradients or parameters, EvoPrompt startsfrom a population of prompts and iteratively generates new prompts with LLMsbased on the evolutionary operators, improving the population based on thedevelopment set. We optimize prompts for both closed- and open-source LLMsincluding GPT-3.5 and Alpaca, on 31 datasets covering language understanding,generation tasks, as well as BIG-Bench Hard (BBH) tasks. EvoPromptsignificantly outperforms human-engineered prompts and existing methods forautomatic prompt generation (e.g., up to 25% on BBH). Furthermore, EvoPromptdemonstrates that connecting LLMs with EAs creates synergies, which couldinspire further research on the combination of LLMs and conventionalalgorithms.

EvoPrompt: Connecting LLMs with Evolutionary Algorithms Yields Powerful Prompt Optimizers