17 papers
Can Released LLM Vocabularies Support Token-Level Estimation of Hidden Corpora?
Qingjie Zhang, Xingzhang Ren, Zixuan Chen +6
Pretraining corpus composition shapes LLM capabilities, but it often remains hidden even when model weights are released. Prior work has inferred corpus mixtures or traced specific…
Auditing Chinese Web-scale Corpora via Sampled BPE Token Statistics
Qingjie Zhang, Ziqi Tang, Jie Zhang +7
Chinese web pollution has surfaced in LLMs, motivating audits of upstream Chinese corpora. However, auditing such corpora faces three challenges: (1) their web-scale size makes ful…
SafeRedir: Prompt Embedding Redirection for Robust Unlearning in Image Generation Models
Renyang Liu, Kangjie Chen, Han Qiu +4
Image generation models (IGMs), while capable of producing impressive and creative content, often memorize a wide range of undesirable concepts from their training data, leading to…
When Search Goes Wrong: Red-Teaming Web-Augmented Large Language Models
Haoran Ou, Kangjie Chen, Xingshuo Han +4
Large Language Models (LLMs) have been augmented with web search to overcome the limitations of the static knowledge boundary by accessing up-to-date information from the open Inte…
State-Dependent Safety Failures in Multi-Turn Language Model Interaction
Pengcheng Li, Jie Zhang, Tianwei Zhang +5
Safety alignment in large language models is typically evaluated under isolated queries, yet real-world use is inherently multi-turn. Although multi-turn jailbreaks are empirically…
BitHydra: Towards Bit-flip Inference Cost Attack against Large Language Models
Xiaobei Yan, Yiming Li, Hao Wang +2
Large language models (LLMs) are widely deployed, but their substantial compute demands make them vulnerable to inference cost attacks that aim to deliberately maximize the output…