Unnatural language processing: how do language models handle machine-generated prompts?

Mostra el registre complet Registre parcial de l'ítem

• dc.contributor.author Kervadec, Corentin
• dc.contributor.author Franzon, Francesca
• dc.contributor.author Baroni, Marco
• dc.date.accessioned 2023-12-18T07:03:09Z
• dc.date.available 2023-12-18T07:03:09Z
• dc.date.issued 2023
• dc.description Comunicació presentada a la Conference on Empirical Methods in Natural Language Processing (EMNLP 2023), celebrada a Singapur del 6 al 10 de desembre de 2023.
• dc.description.abstract Language model prompt optimization research has shown that semantically and grammatically well-formed manually crafted prompts are routinely outperformed by automatically generated token sequences with no apparent meaning or syntactic structure, including sequences of vectors from a model’s embedding space. We use machine-generated prompts to probe how models respond to input that is not composed of natural language expressions. We study the behavior of models of different sizes in multiple semantic tasks in response to both continuous and discrete machine-generated prompts, and compare it to the behavior in response to humangenerated natural-language prompts. Even when producing a similar output, machinegenerated and human prompts trigger different response patterns through the network processing pathways, including different perplexities, different attention and output entropy distributions, and different unit activation profiles. We provide preliminary insight into the nature of the units activated by different prompt types, suggesting that only natural language prompts recruit a genuinely linguistic circuit.
• dc.format.mimetype application/pdf
• dc.identifier.citation Kervadec C, Franzon F, Baroni M. Unnatural language processing: how do language models handle machine-generated prompts?. In: Bouamor H, Pino J, Bali K. Findings of of the 2023 Conference on Empirical Methods in Natural Language Processing; 2023 Dec 6-10; Singapore. East Stroudsburg PA: ACL; 2023. p. 14377-92.
• dc.identifier.isbn 9798891760615
• dc.identifier.uri http://hdl.handle.net/10230/58560
• dc.language.iso eng
• dc.publisher ACL (Association for Computational Linguistics)
• dc.relation.ispartof Findings of of the 2023 Conference on Empirical Methods in Natural Language Processing; 2023 Dec 6-10; Singapore. East Stroudsburg PA: ACL; 2023. p. 14377-92.
• dc.rights © ACL, Creative Commons Attribution 4.0 License
• dc.rights.accessRights info:eu-repo/semantics/openAccess
• dc.rights.uri https://creativecommons.org/licenses/by/4.0/
• dc.subject.other Tractament del llenguatge natural (Informàtica)
• dc.subject.other Llengües artificials
• dc.subject.other Semàntica
• dc.subject.other Lingüística computacional
• dc.title Unnatural language processing: how do language models handle machine-generated prompts?
• dc.type info:eu-repo/semantics/conferenceObject
• dc.type.version info:eu-repo/semantics/publishedVersion