: Vocabulaire du TAL: prompt engineering

NLP methods and tools > artificial intelligence > prompt engineering

prompt engineering

The process of creating natural language instructions, or prompts, to extract knowledge from [language models] in an organized manner. (Vatsal and Dubey. "A Survey of Prompt Engineering Methods in Large Language Models for Different NLP Tasks" (24 Jul 2024). Available at https://doi.org/10.48550/arXiv.2407.12994 (Accessed on 10 Oct 2024))

Despite several attempts at prompt engineering we did not manage to obtain extractive summaries from the model. (Maës, Boudraa, Blache & Becerra-Bonache, 2024)
In the future more detailed prompt engineering can potentially improve the personalization of generated CNs. (Doganç & Markov, 2023)
The prompt engineering required is expensive and timeconsuming as it involves a human manually designing and testing beliefs. (Bauer, Mehrabi, Goyal, Chang, Galstyan & Gupta, 2024)

http://data.loterre.fr/ark:/67375/8LP-KNR5JVVC-Q

RDF/XML TURTLE JSON-LD Dernière modification le 10/10/2024

Loterre