Shaoqing Zhang

Shaoqing Zhang, Zhuosheng Zhang, Kehai Chen et al.

Despite being empowered with alignment mechanisms, large language models (LLMs) are increasingly vulnerable to emerging jailbreak attacks that can compromise their alignment mechanisms. This vulnerability poses significant risks to real-world applications. Existing work faces challenges in both training efficiency and generalization capabilities (i.e., Reinforcement Learning from Human Feedback and Red-Teaming). Developing effective strategies to enable LLMs to resist continuously evolving jailbreak attempts represents a significant challenge. To address this challenge, we propose a novel defensive paradigm called GuidelineLLM, which assists LLMs in recognizing queries that may have harmful content. Before LLMs respond to a query, GuidelineLLM first identifies potential risks associated with the query, summarizes these risks into guideline suggestions, and then feeds these guidelines to the responding LLMs. Importantly, our approach eliminates the necessity for additional safety fine-tuning of the LLMs themselves; only the GuidelineLLM requires fine-tuning. This characteristic enhances the general applicability of GuidelineLLM across various LLMs. Experimental results demonstrate that GuidelineLLM can significantly reduce the attack success rate (ASR) against LLM (an average reduction of 34.17\% ASR) while maintaining the usefulness of LLM in handling benign queries. The code is available at https://github.com/sqzhang-lazy/GuidelineLLM.

Shihan Dou, Jiazheng Zhang, Jianxiang Zang et al.

We introduce MPLSandbox, an out-of-the-box multi-programming language sandbox designed to provide unified and comprehensive feedback from compiler and analysis tools for Large Language Models (LLMs). It can automatically identify the programming language of the code, compiling and executing it within an isolated sub-sandbox to ensure safety and stability. In addition, MPLSandbox also integrates both traditional and LLM-based code analysis tools, providing a comprehensive analysis of generated code. MPLSandbox can be effortlessly integrated into the training and deployment of LLMs to improve the quality and correctness of their generated code. It also helps researchers streamline their workflows for various LLM-based code-related tasks, reducing the development cost. To validate the effectiveness of MPLSandbox, we integrate it into training and deployment approaches, and also employ it to optimize workflows for a wide range of real-world code-related tasks. Our goal is to enhance researcher productivity on LLM-based code-related tasks by simplifying and automating workflows through delegation to MPLSandbox.

Daosheng Xu, Zebin Lu, Jeremy Cheuk-Hin Leung et al.

Given the interpretability, accuracy, and stability of numerical weather prediction (NWP) models, current operational weather forecasting relies heavily on the NWP approach. In the past two years, the rapid development of Artificial Intelligence (AI) has provided an alternative solution for medium-range (1-10 days) weather forecasting. Bi et al. (2023) (hereafter Bi23) introduced the first AI-based weather prediction (AIWP) model in China, named Pangu-Weather, which offers fast prediction without compromising accuracy. In their work, Bi23 made notable claims regarding its effectiveness in extreme weather predictions. However, this claim lacks persuasiveness because the extreme nature of the two tropical cyclones (TCs) examples presented in Bi23, namely Typhoon Kong-rey and Typhoon Yutu, stems primarily from their intensities rather than their moving paths. Their claim may mislead into another meaning which is that Pangu-Weather works well in predicting unusual typhoon paths, which was not explicitly analyzed. Here, we reassess Pangu-Weather's ability to predict extreme TC trajectories from 2020-2024. Results reveal that while Pangu-Weather overall outperforms NWP models in predicting tropical cyclone (TC) tracks, it falls short in accurately predicting the rarely observed sudden-turning tracks, such as Typhoon Khanun in 2023. We argue that current AIWP models still lag behind traditional NWP models in predicting such rare extreme events in medium-range forecasts.

Shaoqing Zhang, Kehai Chen, Zhuosheng Zhang et al.

Recent advancements in vision-language models have increased interest in Device-Control Agents (DC agents) for managing graphical user interfaces (GUIs). With the growing complexity and integration of such agents into various applications, effective evaluation methods have become crucial. The current evaluation method for DC agents primarily focuses on the instruction level, providing the current state (e.g., screenshots) and past execution history to determine actions for target instructions, helping identify potential execution failures. However, in GUI environments, a single state may contain multiple interactive widgets, each linked to different instructions, presenting an opportunity for diverse actions based on various instruction targets. Evaluating the agent's performance solely at the instruction level may overlook the broader context of these interactions. To capture a more comprehensive view of agent performance, we propose a new evaluation method, XBOUND, to evaluate the accuracy of instruction completion on a per-state basis. XBOUND provides a state-level evaluation framework, serving as a tool to assess agents' capabilities within environmental states. Our evaluation yields several key insights: UI-TARS stands out as the strongest 7B model, current agents display a bimodal performance pattern in instruction unification, and sub-7B models remain limited in state mastery. We further identify GPT-based planning as a critical bottleneck, and show that grounding data mainly benefits action matching, while trajectory data is more effective for instruction unification.

Rixu Hao, Yuxin Zhao, Shaoqing Zhang et al.

El Niño-Southern Oscillation (ENSO) exerts global climate and societal impacts, but real-time prediction with lead times beyond one year remains challenging. Dynamical models suffer from large biases and uncertainties, while deep learning struggles with interpretability and multi-scale dynamics. Here, we introduce PTSTnet, an interpretable model that unifies dynamical processes and cross-scale spatiotemporal learning in an innovative neural-network framework with physics-encoding learning. PTSTnet produces interpretable predictions significantly outperforming state-of-the-art benchmarks with lead times beyond 24 months, providing physical insights into error propagation in ocean-atmosphere interactions. PTSTnet learns feature representations with physical consistency from sparse data to tackle inherent multi-scale and multi-physics challenges underlying ocean-atmosphere processes, thereby inherently enhancing long-term prediction skill. Our successful realizations mark substantial steps forward in interpretable insights into innovative neural ocean modelling.