Self-Recognition in Language Models
Paper
•
2407.06946
•
Published
•
23
Note 检测LLM的自我意识,但是没发现证据支持LLMs有自我意识; we propose a novel approach for assessing self-recognition in LMs using model-generated “security questions” it does not require access to internal model parameters or output probabilities. Our extensive experiments found **no empirical evidence of general or consistent self-recognition** in any examined LM.
Note ESFT aims to efficiently customize Large Language Models (LLMs) with Mixture-of-Experts (MoE) architecture by adjusting only task-relevant parts, improving efficiency and performance while using fewer resources and storage. [2R; 持续学习?怎么找到task-relevant parts?]