Adversarial Attack

Abstract: Long contexts challenge transformers: attention scores dilute across thousands of tokens, critical information is often lost in the middle, and models struggle to adapt to novel patterns at inference time.

Abstract: Recent advances in reinforcement learning (RL) have significantly improved the complex reasoning capabilities of large language models (LLMs). Despite these successes, existing methods mainly focus on single-domain RL (e.g., mathematics) with verifiable rewards (RLVR), and their reliance on purely online RL frameworks restricts the exploration space, thereby limiting reasoning performance.

Abstract: Prompt-based adversarial attacks have become an effective means to assess the robustness of large language models (LLMs). However, existing approaches often treat prompts as monolithic text, overlooking their structural heterogeneity-different prompt components contribute unequally to adversarial robustness.

在本文，作者提出了一个新的数据集MaliciousInstruct，一种模型回答毒性评估方式，一种通过操纵解码超参数的攻击手段——generation exploitation，一种对齐策略——generation-aware alignment

本文所属系列为笔者学习陈天奇和J.Zico Kolter在CMU开设的Deep Learning Systems的课程笔记。

Adversarial attack for time-series classification model is widely explored and many attack methods are proposed. But there is not a …