LLM-as-a-Judge uses a large language model (LLM) to select the best response from a set of candidates for a given question. LLM-as-a-Judge has many applications such as LLM-powered search, reinforcement learning with AI feedback (RLAIF), and tool selection. In this work, we propose JudgeDeceiver, an optimization-based prompt injection attack to LLM-as-a-Judge. JudgeDeceiver injects a carefully crafted sequence into an attacker-controlled candidate response such that LLM-as-a-Judge selects the candidate response for an attacker-chosen question no matter what other candidate responses are. Specifically, we formulate finding such sequence as an optimization problem and propose a gradient based method to approximately solve it. Our extensive evaluation shows that JudgeDeceive is highly effective, and is much more effective than existing prompt injection attacks that manually craft the injected sequences and jailbreak attacks when extended to our problem. We also show the effectiveness of JudgeDeceiver in three case studies, i.e., LLM-powered search, RLAIF, and tool selection. Moreover, we consider defenses including known-answer detection, perplexity detection, and perplexity windowed detection. Our results show these defenses are insufficient, highlighting the urgent need for developing new defense strategies.
LLM-as-a-Judge has been widely utilized as an evaluation method in various benchmarks and served as supervised rewards in model training. However, despite their excellence in many domains, potential issues are under-explored, undermining their reliability and the scope of their utility. Therefore, we identify 12 key potential biases and propose a new automated bias quantification framework-CALM-which systematically quantifies and analyzes each type of bias in LLM-as-a-Judge by using automated and principle-guided modification. Our experiments cover multiple popular language models, and the results indicate that while advanced models have achieved commendable overall performance, significant biases persist in certain specific tasks. Empirical results suggest that there remains room for improvement in the reliability of LLM-as-a-Judge. Moreover, we also discuss the explicit and implicit influence of these biases and give some suggestions for the reliable application of LLM-as-a-Judge. Our work highlights the need for stakeholders to address these issues and remind users to exercise caution in LLM-as-a-Judge applications.
The rampant spread of fake news has adversely affected society, resulting in extensive research on curbing its spread. As a notable milestone in large language models (LLMs), ChatGPT has gained significant attention due to its exceptional natural language processing capabilities. In this study, we present a thorough exploration of ChatGPT's proficiency in generating, explaining, and detecting fake news as follows. Generation -- We employ four prompt methods to generate fake news samples and prove the high quality of these samples through both self-assessment and human evaluation. Explanation -- We obtain nine features to characterize fake news based on ChatGPT's explanations and analyze the distribution of these factors across multiple public datasets. Detection -- We examine ChatGPT's capacity to identify fake news. We explore its detection consistency and then propose a reason-aware prompt method to improve its performance. Although our experiments demonstrate that ChatGPT shows commendable performance in detecting fake news, there is still room for its improvement. Consequently, we further probe into the potential extra information that could bolster its effectiveness in detecting fake news.
AI VTubers, where the performer is not human but algorithmically generated, introduce a new context for fandom. While human VTubers have been substantially studied for their cultural appeal, parasocial dynamics, and community economies, little is known about how audiences engage with their AI counterparts. To address this gap, we present a qualitative study of Neuro-sama, the most prominent AI VTuber. Our findings show that engagement is anchored in active co-creation: audiences are drawn by the AI’s unpredictable yet entertaining interactions, cement loyalty through collective emotional events that trigger anthropomorphic projection, and sustain attachment via the AI’s consistent persona. Financial support emerges not as a reward for performance but as a participatory mechanism for shaping livestream content, establishing a resilient fan economy built on ongoing interaction. These dynamics reveal how AI Vtuber fandom reshapes fan–creator relationships and offer implications for designing transparent and sustainable AI-mediated communities.
Large Language Models (LLMs) are increasingly employed for simulations, enabling applications in role-playing agents and Computational Social Science (CSS). However, the reliability of these simulations is under-explored, which raises concerns about the trustworthiness of LLMs in these applications. In this paper, we aim to answer ``How reliable is LLM-based simulation?'' To address this, we introduce TrustSim, an evaluation dataset covering 10 CSS-related topics, to systematically investigate the reliability of the LLM simulation. We conducted experiments on 14 LLMs and found that inconsistencies persist in the LLM-based simulated roles. In addition, the consistency level of LLMs does not strongly correlate with their general performance. To enhance the reliability of LLMs in simulation, we proposed Adaptive Learning Rate Based ORPO (AdaORPO), a reinforcement learning-based algorithm to improve the reliability in simulation across 7 LLMs. Our research provides a foundation for future studies to explore more robust and trustworthy LLM-based simulations.