LIONs: An Empirically Optimized Approach to Align Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913528714100736 |
|---|---|
| author | Yu, Xiao Wu, Qingyang Li, Yu Yu, Zhou |
| author_facet | Yu, Xiao Wu, Qingyang Li, Yu Yu, Zhou |
| contents | Alignment is a crucial step to enhance the instruction-following and conversational abilities of language models. Despite many recent work proposing new algorithms, datasets, and training pipelines, there is a lack of comprehensive studies measuring the impact of various design choices throughout the whole training process. We first conduct a rigorous analysis over a three-stage training pipeline consisting of supervised fine-tuning, offline preference learning, and online preference learning. We have found that using techniques like sequence packing, loss masking in SFT, increasing the preference dataset size in DPO, and online DPO training can significantly improve the performance of language models. We then train from Gemma-2b-base and LLama-3-8b-base, and find that our best models exceed the performance of the official instruct models tuned with closed-source data and algorithms. Our code and models can be found at \url{https://github.com/Columbia-NLP-Lab/LionAlignment}. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_06542 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | LIONs: An Empirically Optimized Approach to Align Language Models Yu, Xiao Wu, Qingyang Li, Yu Yu, Zhou Computation and Language Alignment is a crucial step to enhance the instruction-following and conversational abilities of language models. Despite many recent work proposing new algorithms, datasets, and training pipelines, there is a lack of comprehensive studies measuring the impact of various design choices throughout the whole training process. We first conduct a rigorous analysis over a three-stage training pipeline consisting of supervised fine-tuning, offline preference learning, and online preference learning. We have found that using techniques like sequence packing, loss masking in SFT, increasing the preference dataset size in DPO, and online DPO training can significantly improve the performance of language models. We then train from Gemma-2b-base and LLama-3-8b-base, and find that our best models exceed the performance of the official instruct models tuned with closed-source data and algorithms. Our code and models can be found at \url{https://github.com/Columbia-NLP-Lab/LionAlignment}. |
| title | LIONs: An Empirically Optimized Approach to Align Language Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2407.06542 |