Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Singh, Vaibhav, Krishna, Amrith, NJ, Karthika, Ramakrishnan, Ganesh
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2406.17377
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909231106490368
author Singh, Vaibhav
Krishna, Amrith
NJ, Karthika
Ramakrishnan, Ganesh
author_facet Singh, Vaibhav
Krishna, Amrith
NJ, Karthika
Ramakrishnan, Ganesh
contents Low-resource languages, by its very definition, tend to be under represented in the pre-training corpora of Large Language Models. In this work, we investigate three low-resource cross-lingual approaches that enable an LLM adapt to tasks in previously unseen languages. Llama-2 is an LLM where Indic languages, among many other language families, contribute to less than $0.005\%$ of the total $2$ trillion token pre-training corpora. In this work, we experiment with the English-dominated Llama-2 for cross-lingual transfer to three Indic languages, Bengali, Hindi, and Tamil as target languages. We study three approaches for cross-lingual transfer, under ICL and fine-tuning. One, we find that adding additional supervisory signals via a dominant language in the LLM, leads to improvements, both under in-context learning and fine-tuning. Two, adapting the target languages to word reordering may be beneficial under ICL, but its impact diminishes with fine tuning. Finally, continued pre-training in one low-resource language can improve model performance for other related low-resource languages.
format Preprint
id arxiv_https___arxiv_org_abs_2406_17377
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Three-Pronged Approach to Cross-Lingual Adaptation with Multilingual LLMs
Singh, Vaibhav
Krishna, Amrith
NJ, Karthika
Ramakrishnan, Ganesh
Computation and Language
Low-resource languages, by its very definition, tend to be under represented in the pre-training corpora of Large Language Models. In this work, we investigate three low-resource cross-lingual approaches that enable an LLM adapt to tasks in previously unseen languages. Llama-2 is an LLM where Indic languages, among many other language families, contribute to less than $0.005\%$ of the total $2$ trillion token pre-training corpora. In this work, we experiment with the English-dominated Llama-2 for cross-lingual transfer to three Indic languages, Bengali, Hindi, and Tamil as target languages. We study three approaches for cross-lingual transfer, under ICL and fine-tuning. One, we find that adding additional supervisory signals via a dominant language in the LLM, leads to improvements, both under in-context learning and fine-tuning. Two, adapting the target languages to word reordering may be beneficial under ICL, but its impact diminishes with fine tuning. Finally, continued pre-training in one low-resource language can improve model performance for other related low-resource languages.
title A Three-Pronged Approach to Cross-Lingual Adaptation with Multilingual LLMs
topic Computation and Language
url https://arxiv.org/abs/2406.17377