Contrastive Post-training Large Language Models on Data Curriculum
Keywords: large language model, curriculum learning, contrastive learning, alignment
TL;DR: Contrastive post-training can improve LLM alignment without feedback from human or AI.
Abstract: Alignment serves as an important step to steer large language models (LLMs) towards human preferences. In this paper, we explore contrastive post-training techniques for alignment by automatically constructing preference pairs from multiple models of varying strengths (e.g., InstructGPT, ChatGPT and GPT-4). We carefully compare the contrastive techniques of SLiC and DPO to SFT baselines and find that DPO provides a step-function improvement even after continueing SFT saturates. We also explore a data curriculum learning scheme for contrastive post-training, which starts by learning from "easier" pairs and transitioning to "harder" ones, which further improves alignment. Finally, we scale up our experiments to train with more data and larger models like Orca. Remarkably, contrastive post-training further improves the performance of Orca, already a state-of-the-art instruction learning model tuned with GPT-4 outputs, to exceed that of ChatGPT.
Primary Area: representation learning for computer vision, audio, language, and other modalities
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 8492
Loading