Advanced LLMs have achieved near-ceiling instruction-following accuracy on benchmarks such as IFEVAL.However, these impressive scores do not necessarily translate to reliable services in real-world use, where users often vary their phrasing, contextual framing, and task formulations.In this paper, we study nuance-oriented reliability: whether models exhibit consistent competence across cousin prompts that convey analogous user intents but with subtle nuances.To quantify this, we introduce a new metric, reliable@k, and develop an automated pipeline that generates high-quality cousin prompts via data augmentation.Building upon this, we construct IFE-VAL++ for systematic evaluation.Across 20 proprietary and 26 open-source LLMs, we find that current models exhibit substantial insufficiency in nuance-oriented reliability-their performance can drop by up to 61.8% with nuanced prompt modifications.What's more, we characterize it and explore three potential improvement recipes.Our findings highlight nuance-oriented reliability as a crucial yet underexplored next step toward more dependable and trustworthy LLM behavior.Our code and benchmark are accessible: https: //github.com/jianshuod/IFEval-pp.

Qingjie Zhang, Di Wang, Haoting Qian, Yiming Li, Tianwei Zhang, Minlie Huang, Ke Xu, Hewu Li, Liu Yan, Han Qiu. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Large language models (LLMs) exhibit prompt leakage vulnerabilities, where they may be coaxed into revealing system prompts embedded in LLM services, raising intellectual property and confidentiality concerns.An intriguing question arises: Do LLMs genuinely internalize prompt leakage intents in their hidden states before generating tokens?In this work, we use probing techniques to capture LLMs' intent-related internal representations and confirm that the answer is yes.We start by comprehensively inducing prompt leakage behaviors across diverse system prompts, attack queries, and decoding methods.We develop a hybrid labeling pipeline, enabling the identification of broader prompt leakage behaviors beyond mere verbatim leaks.Our results show that a simple linear probe can predict prompt leakage risks from pre-generation hidden states without generating any tokens.Across all tested models, linear probes consistently achieve 90%+ AUROC, even when applied to new system prompts and attacks.Understanding the model internals behind prompt leakage drives practical applications, including intention-based detection of prompt leakage risks.

Qingjie Zhang, Di Wang, Haoting Qian, Liu Yan, Tianwei Zhang, Ke Xu, Qi Li, Minlie Huang, Hewu Li, Han Qiu. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Large language models (LLMs) can output sensitive information, which has emerged as a novel safety concern. Previous works focus on structured sensitive information (e.g. personal identifiable information). However, we notice that sensitive information can also be at semantic level, i.e. semantic sensitive information (SemSI). Particularly, *simple natural questions* can let state-of-the-art (SOTA) LLMs output SemSI. %which is hard to be detected compared with structured ones. Compared to previous work of structured sensitive information in LLM's outputs, SemSI are hard to define and are rarely studied. Therefore, we propose a novel and large-scale investigation on the existence of SemSI in SOTA LLMs induced by simple natural questions. First, we construct a comprehensive and labeled dataset of semantic sensitive information, SemSI-Set, by including three typical categories of SemSI. Then, we propose a large-scale benchmark, SemSI-Bench, to systematically evaluate semantic sensitive information in 25 SOTA LLMs. Our finding reveals that SemSI widely exists in SOTA LLMs' outputs by querying with simple natural questions. We open-source our project at https://semsi-project.github.io/.

Name lookup is a key technology for the forwarding plane of content router in Named Data Networking (NDN). To realize the efficient name lookup, what counts is deploying a high-performance index in content routers. So far, the proposed indexes have shown good performance, most of which are optimized for or evaluated with URLs collected from the current Internet, as the large-scale NDN names are not available yet. Unfortunately, the performance of these indexes is always impacted in terms of lookup speed, memory consumption and false positive probability, as the distributions of URLs retrieved in memory may differ from those of real NDN names independently generated by content-centric applications online. Focusing on this gap, a smart mapping model named Pyramid-NN via neural networks is proposed to build an index called LNI for NDN forwarding plane. Through learning the distributions of the names retrieved in the static memory, LNI that will be trained by real NDN names offline and preset in content routers in the future can not only reduce the memory consumption and the probability of false positive, but also ensure the performance of real NDN name lookup. Experimental results show that LNI-based FIB can reduce the memory consumption to 58.258 MB. Moreover, as it can be deployed on SRAMs, the throughput is about 177 MSPS, which well meets the current network requirement for fast packet processing.

ABSTRACT A base station is the controller and the data‐receiving center of a wireless sensor network. Hence, a reliable and secure base station is critical to the network. Once an attacker locates the base station, he or she can do many damages to the network. In this paper, we examine the base station location privacy problem from both the attack and defense sides. First, we present a new attack on base station: parent‐based attack scheme (PAS). PAS can locate a base station within one radio (wireless transmission) range of sensors in high‐density sensor networks. Different from existing methods, PAS determines the base station location on the basis of parent–child relationship of sensor nodes. Existing base station protection schemes cannot defend against PAS. Second, on the basis of PAS, we propose a two‐phase parent‐based attack scheme (TP‐PAS). Our simulation results demonstrate that TP‐PAS is able to determine the base station successfully in both low‐density and high‐density sensor networks. Then, to defend against PAS and TP‐PAS, we design a child‐based routing protocol and a parent‐free routing protocol for sensor networks. Our theory analysis and experiment results show that the parent‐free routing protocol has more communication cost and less end‐to‐end latency compared with the child‐based routing protocol. Copyright © 2012 John Wiley & Sons, Ltd.

This paper describes a novel approach to the presentation of results from Web-search by utilizing the semantic correlations amongst the retrieved results as well as their spatio-temporal characteristics. The fundamental goals of this research are threefold; first, it seeks to facilitate easier access to information than what is provided by the current paradigm of listing the retrieved hits in order of relevance. Second, and more importantly, it attempts to recognize, retain, and present context within the retrieved data. Such support is essential for assimilation of the information obtained as the consequence of a query. Finally, it seeks to support both direct search behavior as well as more exploratory search strategies. In this paper, we describe the algorithmic underpinnings, based on latent semantic analysis, of our approach as well as the strategy for spatio-temporal information display and querying. Examples involving different types of queries and comparative experiments illustrate the idea and verify its efficacy