研究人员开发了用于解释机器学习模型内部工作原理的新方法。一种方法是在冻结的语言模型上训练轻量级适配器，以实现可靠的自解释，从而提高主题识别和隐式推理等任务的性能。另一种方法 IdEst 使用内在维度估计来评估自监督学习表示，与下游性能高度相关并支持高效的超参数调整。第三篇论文介绍了 KREPES，一个使用表示符地标和 Nyström 近似来解析 SSL 表示的框架，揭示算法偏差并实现可扩展分析。 AI
影响 这些可解释性方面的进步可能带来更值得信赖和更易于理解的 AI 系统，有助于调试和偏差检测。
排序理由 多篇 arXiv 学术论文发表，详细介绍了 AI 可解释性和表示评估方面的新研究。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2602.10352v2 Announce Type: replace-cross Abstract: Self-interpretation methods prompt language models to describe their own internal states, but remain unreliable due to hyperparameter sensitivity. We show that training lightweight adapters on interpretability artifacts, w…
arXiv:2606.03338v1 Announce Type: new Abstract: Self-supervised learning (SSL) has emerged as a powerful paradigm for learning meaningful representations from unlabeled data. However, the standard protocol for evaluating these representations, linear probing, is computationally e…
arXiv:2509.24467v3 Announce Type: replace-cross Abstract: Self-supervised learning (SSL) learns representations from massive unlabeled data, yet the resulting models typically operate as black boxes, necessitating domain-specific explanations. We introduce KREPES, a unified frame…