Contrastive Instruction Tuning

Tianyi Yan, Fei Wang, James Y. Huang, Wenxuan Zhou, Fan Yin, Aram Galstyan, Wenpeng Yin, Muhao Chen


Abstract
Instruction tuning has been used as a promising approach to improve the performance of large language models (LLMs) on unseen tasks. However, current LLMs exhibit limited robustness to unseen instructions, generating inconsistent outputs when the same instruction is phrased with slightly varied forms or language styles. This behavior indicates LLMs’ lack of robustness to textual variations and generalizability to unseen instructions, potentially leading to trustworthiness issues. Accordingly, we propose Contrastive Instruction Tuning, which maximizes the similarity between the hidden representations of semantically equivalent instruction-instance pairs while minimizing the similarity between semantically different ones. To facilitate this approach, we augment the existing FLAN collection by paraphrasing task instructions. Experiments on the PromptBench benchmark show that CoIN consistently improves LLMs’ robustness to unseen instructions with variations across character, word, sentence, and semantic levels by an average of +2.5% in accuracy.
Anthology ID:
2024.findings-acl.613
Volume:
Findings of the Association for Computational Linguistics: ACL 2024
Month:
August
Year:
2024
Address:
Bangkok, Thailand
Editors:
Lun-Wei Ku, Andre Martins, Vivek Srikumar
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
10288–10302
Language:
URL:
https://aclanthology.org/2024.findings-acl.613
DOI:
10.18653/v1/2024.findings-acl.613
Bibkey:
Cite (ACL):
Tianyi Yan, Fei Wang, James Y. Huang, Wenxuan Zhou, Fan Yin, Aram Galstyan, Wenpeng Yin, and Muhao Chen. 2024. Contrastive Instruction Tuning. In Findings of the Association for Computational Linguistics: ACL 2024, pages 10288–10302, Bangkok, Thailand. Association for Computational Linguistics.
Cite (Informal):
Contrastive Instruction Tuning (Yan et al., Findings 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.findings-acl.613.pdf