Summary of Advancing Parameter Efficiency in Fine-tuning Via Representation Editing, by Muling Wu et al.
Advancing Parameter Efficiency in Fine-tuning via Representation Editing
by Muling Wu, Wenhao Liu, Xiaohua Wang, Tianlong Li, Changze Lv, Zixuan Ling, Jianhao Zhu, Cenyuan Zhang, Xiaoqing Zheng, Xuanjing Huang
First submitted to arxiv on: 23 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computation and Language (cs.CL)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel fine-tuning approach called Representation EDiting (RED), which reduces the number of trainable parameters in neural models by up to 32 times compared to traditional full parameter fine-tuning. RED achieves competitive results with existing Parameter Efficient Fine-Tuning (PEFT) methods, such as LoRA and Adapter, while requiring minimal hyperparameter tuning. The method is tested on various model architectures, including RoBERTa, GPT-2, T5, and LLaMA-2, demonstrating its effectiveness and efficiency. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary RED is a simple yet powerful approach that modifies the representations generated by some layers in neural models through scaling and biasing operations. This technique can significantly reduce the number of trainable parameters without sacrificing performance, making it an attractive option for large-scale neural models. |
Keywords
* Artificial intelligence * Fine tuning * Gpt * Hyperparameter * Llama * Lora * Parameter efficient * T5
