paper-with-me

홈 › Papers

SymNoise: Advancing Language Model Fine-tuning with Symmetric Noise

2023-12-03 · Abhay Kumar Yadav, Arjun Singh

In this paper, we introduce a novel fine-tuning technique for language models, which involves incorporating symmetric noise into the embedding process. This method aims to enhance the model's function by more stringently regulating its local curvature, demonstrating superior performance over the current method, NEFTune. When fine-tuning the LLaMA-2-7B model using Alpaca, standard techniques yield a 29.79% score on AlpacaEval. However, our approach, SymNoise, increases this score significantly to 69.04%, using symmetric noisy embeddings. This is a 6.7% improvement over the state-of-the-art method, NEFTune~(64.69%). Furthermore, when tested on various models and stronger baseline instruction datasets, such as Evol-Instruct, ShareGPT, OpenPlatypus, SymNoise consistently outperforms NEFTune. The current literature, including NEFTune, has underscored the importance of more in-depth research into the application of noise-based strategies in the fine-tuning of language models. Our approach, SymNoise, is another significant step towards this direction, showing notable improvement over the existing state-of-the-art method.

📄 PDF Abstract BibTeX arXiv:2312.01523

Code (0)

등록된 구현이 없습니다.

Tasks

Language ModelingLanguage Modelling

Similar Papers 제목 키워드 기반

Understanding and Improving Noisy Embedding Techniques in Instruction Finetuning

2026-05-22 · Abhay Yadav arxiv

Recent advancements in instructional fine-tuning have injected noise into embeddings, with NEFTune (Jain et al., 2024) setting benchmarks using uniform noise. Despite NEFTune's empirical findings that uniform noise outpe…

RLIPv2: Fast Scaling of Relational Language-Image Pre-training

2023-08-18 · ICCV 2023 1 · Hangjie Yuan, Shiwei Zhang, Xiang Wang, Samuel Albanie 외

Relational Language-Image Pre-training (RLIP) aims to align vision representations with relational texts, thereby advancing the capability of relational reasoning in computer vision tasks. However, hindered by the slow c…

Graph GenerationHuman-Object Interaction Detectionobject-detectionObject Detection+4

Donors and Recipients: On Asymmetric Transfer Across Tasks and Languages with Parameter-Efficient Fine-Tuning

2025-11-17 · Kajetan Dymkiewicz, Ivan Vulic, Helen Yannakoudakis, Eilam Shapira 외 arxiv

Large language models (LLMs) perform strongly across tasks and languages, yet how improvements in one task or language affect other tasks and languages remains poorly understood. We conduct a controlled LoRA fine-tuning …

parameter-efficient fine-tuning

VLA-Arena: An Open-Source Framework for Benchmarking Vision-Language-Action Models

2025-12-27 · Borong Zhang, Jiahao Li, Jiachen Shen, Yuhao Zhang 외 arxiv

While Vision-Language-Action models (VLAs) are rapidly advancing towards generalist robot policies, it remains difficult to quantitatively understand their limits and failure modes. To address this, we introduce a compre…

HydraLoRA: An Asymmetric LoRA Architecture for Efficient Fine-Tuning

2024-04-30 · Chunlin Tian, Zhan Shi, Zhijiang Guo, Li Li 외

Adapting Large Language Models (LLMs) to new tasks through fine-tuning has been made more efficient by the introduction of Parameter-Efficient Fine-Tuning (PEFT) techniques, such as LoRA. However, these methods often und…

parameter-efficient fine-tuning