arXiv Open Access 2024

The Impact of Copyrighted Material on Large Language Models: A Norwegian Perspective

Javier de la Rosa Vladislav Mikhailov Lemei Zhang Freddy Wetjen David Samuel +14 lainnya
Lihat Sumber

Abstrak

The use of copyrighted materials in training language models raises critical legal and ethical questions. This paper presents a framework for and the results of empirically assessing the impact of publisher-controlled copyrighted corpora on the performance of generative large language models (LLMs) for Norwegian. When evaluated on a diverse set of tasks, we found that adding both books and newspapers to the data mixture of LLMs tend to improve their performance, while the addition of fiction works seems to be detrimental. Our experiments could inform the creation of a compensation scheme for authors whose works contribute to AI development.

Topik & Kata Kunci

Penulis (19)

J

Javier de la Rosa

V

Vladislav Mikhailov

L

Lemei Zhang

F

Freddy Wetjen

D

David Samuel

P

Peng Liu

R

Rolv-Arild Braaten

P

Petter Mæhlum

M

Magnus Breder Birkenes

A

Andrey Kutuzov

T

Tita Enstad

H

Hans Christian Farsethås

S

Svein Arne Brygfjeld

J

Jon Atle Gulla

S

Stephan Oepen

E

Erik Velldal

W

Wilfred Østgulen

L

Liljia Øvrelid

A

Aslak Sira Myhre

Format Sitasi

Rosa, J.d.l., Mikhailov, V., Zhang, L., Wetjen, F., Samuel, D., Liu, P. et al. (2024). The Impact of Copyrighted Material on Large Language Models: A Norwegian Perspective. https://arxiv.org/abs/2412.09460

Akses Cepat

Lihat di Sumber
Informasi Jurnal
Tahun Terbit
2024
Bahasa
en
Sumber Database
arXiv
Akses
Open Access ✓