边端语言模型易产生偏见，本文提出反馈机制有效降低偏见。
Biases in Edge Language Models: Detection, Analysis, and Mitigation
- 在树莓派上运行的Llama-2比桌面和云端模型更易随时间产生偏见。
- 在树莓派上部署的Llama-2偏见程度比桌面环境高43.23%。
- 通过分层约束反馈机制，可减少79.28%的模型偏见，适合边缘AI伦理研究者。
将大语言模型（LLM）部署于树莓派等低功耗设备上形成的边端语言模型（ELM），为个性化、安全且低延迟的语言智能提供了可能。然而，设备资源受限及模型缺乏稳健的伦理保障，导致输出公平性、可问责性与透明性面临挑战。本文对比分析了在边端、云和桌面环境中部署的语言模型在文本偏见方面的表现，考察部署环境对模型公平性的影响。具体测试包括在树莓派4上运行的优化版Llama-2，以及在云服务器上运行的GPT-4o-mini、Gemini-1.5-flash和Grok-beta，还有在MacOS桌面机上运行的Gemma2和Mistral。结果显示，运行于树莓派4的Llama-2在时序上分别比桌面和云环境下的模型偏见程度高出43.23%和21.89%。为此，本文提出一种反馈循环机制：在推理过程中逐层应用预设的约束权重，根据历史输出动态调整模型行为，实现偏见模式修正，最终使模型偏见降低79.28%。
原文摘要 · Abstract (English)
The integration of large language models (LLMs) on low-power edge devices such as Raspberry Pi, known as edge language models (ELMs), has introduced opportunities for more personalized, secure, and low-latency language intelligence that is accessible to all. However, the resource constraints inherent in edge devices and the lack of robust ethical safeguards in language models raise significant concerns about fairness, accountability, and transparency in model output generation. This paper conducts a comparative analysis of text-based bias across language model deployments on edge, cloud, and desktop environments, aiming to evaluate how deployment settings influence model fairness. Specifically, we examined an optimized Llama-2 model running on a Raspberry Pi 4; GPT 4o-mini, Gemini-1.5-flash, and Grok-beta models running on cloud servers; and Gemma2 and Mistral models running on a MacOS desktop machine. Our results demonstrate that Llama-2 running on Raspberry Pi 4 is 43.23% and 21.89% more prone to showing bias over time compared to models running on the desktop and cloud-based environments. We also propose the implementation of a feedback loop, a mechanism that iteratively adjusts model behavior based on previous outputs, where predefined constraint weights are applied layer-by-layer during inference, allowing the model to correct bias patterns, resulting in 79.28% reduction in model bias.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。