The safety capabilities of large language models (LLMs) are highly uneven across different languages, often allowing cross-lingual attacks to bypass safety guardrails. Existing alignment methods predominantly rely on external alignment signals and provide limited insight into underlying neuron-level mechanisms, making...
Cheng-Yu Xie, Fei Shen, Shu-Yi Miao et al.· 0 citations
Over-refusal is a critical side effect of safety alignment in large language models (LLMs), causing harmless but superficially risky queries to be mistakenly rejected. Despite the effectiveness of existing mitigation methods, why safety-aligned LLMs may still reject harmless queries under correct safety supervision rem...
Chao Chen, Pengyang Shao, Wei Qin et al.· 0 citations
Long-term memory enables LLM agents to learn from past interactions but exposes memory formation to adversarial influence. Existing studies primarily organize memory poisoning by attack channels, corruption patterns, or lifecycle stages, leaving formation interfaces insufficiently characterized. We formulate Memory For...
Haowen Pan, Pengyang Shao, Yun-Shan Ma et al.· 0 citations
Uncovering the internal mechanisms underlying the safety capabilities of large language models (LLMs) is crucial for developing trustworthy artificial intelligence. Currently, mechanistic interpretability studies on multilingual safety are largely confined to local components, such as isolated neurons. However, this st...
Shu-Yi Miao, Wang-Jie Qiu, Pengyang Shao et al.· 0 citations
This work proposes aligning cross-lingual thoughts and responses (ACTR), a framework that improves multilingual safety alignment by strengthening the use of existing safety reasoning, and devise neuron-selective consistency optimization (NSCO), which uses a frozen judge model to reward agreement between the safety cate...
Xian-Hui Zhang, Jian Yu, Cheng-Yu Xie et al.· 0 citations
An agent harness, the code that turns a model into an agent, writes its own record of each run, and that record is all a later reader gets when a run is disputed, investigated or audited. We call a record evidentiary when a reader who was not there can check it without trusting the writer. Across sixteen deployed frame...
Jia-Hong Dai, Zhuo-Chen Yang, Pengyang Shao et al.· 1 citation
Large language models have shown promising potential in chemistry, with prior work exploring molecular recognition, classification, and property prediction. Despite the achieved progress, LLMs are still far from satisfactory when dealing with complex chemical multi-step reaction pathway inference task due to the lack o...
Fan Yang, Fei-Yang Xu, Kun Zhang et al.· Proceedings of the Thirty-Fi...· 0 citations
The workshop brings together researchers and practitioners from data mining, LLMs, NLP, NLP, IR, human-centered AI, and AI safety to position personalization as a central research direction for next-generation AI systems at KDD.
Xiaoyan Zhao, Yang Zhang, Mo-Xin Li et al.· Proceedings of the 32nd ACM...· 0 citations
CLLPU (Cross-Lingual and Language-Bound Protocol for LLM Unlearning), a multilingual benchmark that formulates this problem through two settings: common-goal forgetting, where target knowledge should be suppressed across all languages, and language-conditioned forgetting, where suppression should remain confined to a d...
Pengyang Shao, Chuan-Peng Lu, Wei Qin et al.· 2 citations
Large language models (LLMs) and agentic AI systems are rapidly moving into user-facing applications, yet most remain fundamentally generic, optimized for population-level objectives under the assumption that one model can serve all users. This assumption is increasingly misaligned with real-world deployment, where AI...
Xiaoyan Zhao, Yang Zhang, Moxin Li et al.· Proceedings of the 32nd ACM...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
