Layer by Layer: Uncovering Where Multi-Task Learning Happens in Instruction-Tuned Large Language Models
Open Access
- 1 January 2024
- conference paper
- Published by Association for Computational Linguistics (ACL)
- p. 15195-15214
Abstract
Zheng Zhao, Yftah Ziser, Shay B Cohen. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.Keywords
This publication has 15 references indexed in Scilit:
- LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models2024
- A Joint Matrix Factorization Analysis of Multilingual Representations2023
- Understanding Domain Learning in Language Models Through Subpopulation Analysis2022
- Cross-Task Generalization via Natural Language Crowdsourcing Instructions2022
- Muppet: Massive Multi-task Representations with Pre-Finetuning2021
- First Align, then Predict: Understanding the Cross-Lingual Ability of Multilingual BERT2021
- A Primer in BERTology: What We Know About How BERT WorksTransactions of the Association for Computational Linguistics, 2020
- What Happens To BERT Embeddings During Fine-tuning?2020
- Transformers: State-of-the-Art Natural Language Processing2020
- Multitask LearningMachine Learning, 1997