Adversarial Preference Learning for Robust LLM Alignment

Modern language models often rely on Reinforcement Learning from Human Feedback (RLHF) to encourage safe behaviors. However, they remain vulnerable to adversarial attacks due to three key limitations: (1) the inefficiency and high cost of human annotation, (2) the vast diversity of potential adversarial attacks, and (3) the risk of feedback bias and reward hacking. To address these challenges, we introduce Adversarial Preference Learning (APL), an iterative adversarial training method incorporating three key innovations. First, a direct harmfulness metric based on the model's intrinsic preference probabilities, eliminating reliance on external assessment. Second, a conditional generative attacker that synthesizes input-specific adversarial variations. Third, an iterative framework with automated closed-loop feedback, enabling continuous adaptation through vulnerability discovery and mitigation. Experiments on Mistral-7B-Instruct-v0.3 demonstrate that APL significantly enhances robustness, achieving 83.33% harmlessness win rate over the base model (evaluated by GPT-4o), reducing harmful outputs from 5.88% to 0.43% (measured by LLaMA-Guard), and lowering attack success rate by up to 65% according to HarmBench. Notably, APL maintains competitive utility, with an MT-Bench score of 6.59 (comparable to the baseline 6.78) and an LC-WinRate of 46.52% against the base model.

SurveyX: Academic Survey Automation via Large Language Models

Large Language Models (LLMs) have demonstrated exceptional comprehension capabilities and a vast knowledge base, suggesting that LLMs can serve as efficient tools for automated survey generation. However, recent research related to automated survey generation remains constrained by some critical limitations like finite context window, lack of in-depth content discussion, and absence of systematic evaluation frameworks. Inspired by human writing processes, we propose SurveyX, an efficient and organized system for automated survey generation that decomposes the survey composing process into two phases: the Preparation and Generation phases. By innovatively introducing online reference retrieval, a pre-processing method called AttributeTree, and a re-polishing process, SurveyX significantly enhances the efficacy of survey composition. Experimental evaluation results show that SurveyX outperforms existing automated survey generation systems in content quality (0.259 improvement) and citation quality (1.76 enhancement), approaching human expert performance across multiple evaluation dimensions. Examples of surveys generated by SurveyX are available on www.surveyx.cn

NewsBench: Systematic Evaluation of LLMs for Writing Proficiency and Safety Adherence in Chinese Journalistic Editorial Applications

This study presents NewsBench, a novel benchmark framework developed to evaluate the capability of Large Language Models (LLMs) in Chinese Journalistic Writing Proficiency (JWP) and their Safety Adherence (SA), addressing the gap between journalistic ethics and the risks associated with AI utilization. Comprising 1,267 tasks across 5 editorial applications, 7 aspects (including safety and journalistic writing with 4 detailed facets), and spanning 24 news topics domains, NewsBench employs two GPT-4 based automatic evaluation protocols validated by human assessment. Our comprehensive analysis of 11 LLMs highlighted GPT-4 and ERNIE Bot as top performers, yet revealed a relative deficiency in journalistic ethic adherence during creative writing tasks. These findings underscore the need for enhanced ethical guidance in AI-generated journalistic content, marking a step forward in aligning AI capabilities with journalistic standards and safety considerations.

LCAUnet: A skin lesion segmentation network with enhanced edge and body fusion

Accurate segmentation of skin lesions in dermatoscopic images is crucial for the early diagnosis of skin cancer and improving the survival rate of patients. However, it is still a challenging task due to the irregularity of lesion areas, the fuzziness of boundaries, and other complex interference factors. In this paper, a novel LCAUnet is proposed to improve the ability of complementary representation with fusion of edge and body features, which are often paid little attentions in traditional methods. First, two separate branches are set for edge and body segmentation with CNNs and Transformer based architecture respectively. Then, LCAF module is utilized to fuse feature maps of edge and body of the same level by local cross-attention operation in encoder stage. Furthermore, PGMF module is embedded for feature integration with prior guided multi-scale adaption. Comprehensive experiments on public available dataset ISIC 2017, ISIC 2018, and PH2 demonstrate that LCAUnet outperforms most state-of-the-art methods. The ablation studies also verify the effectiveness of the proposed fusion techniques.