Jialong Li

Jinyu Cai, Jialong Li, Mingyue Zhang et al.

Social media platforms such as Twitter, Reddit, and Sina Weibo play a crucial role in global communication but often encounter strict regulations in geopolitically sensitive regions. This situation has prompted users to ingeniously modify their way of communicating, frequently resorting to coded language in these regulated social media environments. This shift in communication is not merely a strategy to counteract regulation, but a vivid manifestation of language evolution, demonstrating how language naturally evolves under societal and technological pressures. Studying the evolution of language in regulated social media contexts is of significant importance for ensuring freedom of speech, optimizing content moderation, and advancing linguistic research. This paper proposes a multi-agent simulation framework using Large Language Models (LLMs) to explore the evolution of user language in regulated social media environments. The framework employs LLM-driven agents: supervisory agent who enforce dialogue supervision and participant agents who evolve their language strategies while engaging in conversation, simulating the evolution of communication styles under strict regulations aimed at evading social media regulation. The study evaluates the framework's effectiveness through a range of scenarios from abstract scenarios to real-world situations. Key findings indicate that LLMs are capable of simulating nuanced language dynamics and interactions in constrained settings, showing improvement in both evading supervision and information accuracy as evolution progresses. Furthermore, it was found that LLM agents adopt different strategies for different scenarios.

Haozheng Luo, Jiahao Yu, Wenxin Zhang et al.

We introduce a low-resource safety enhancement method for aligning large language models (LLMs) without the need for supervised fine-tuning (SFT) or reinforcement learning from human feedback (RLHF). Our main idea is to exploit knowledge distillation to extract the alignment information from existing well-aligned LLMs and integrate it into unaligned LLMs in a plug-and-play fashion. Methodology, we employ delta debugging to identify the critical components of knowledge necessary for effective distillation. On the harmful question dataset, our method significantly enhances the average defense success rate by approximately 14.41%, reaching as high as 51.39%, in 17 unaligned pre-trained LLMs, without compromising performance.

Jinyu Cai, Jialong Li, Mingyue Zhang et al.

In this paper, we address the challenges faced by Value Iteration Networks (VIN) in handling larger input maps and mitigating the impact of accumulated errors caused by increased iterations. We propose a novel approach, Value Iteration Networks with Gated Summarization Module (GS-VIN), which incorporates two main improvements: (1) employing an Adaptive Iteration Strategy in the Value Iteration module to reduce the number of iterations, and (2) introducing a Gated Summarization module to summarize the iterative process. The adaptive iteration strategy uses larger convolution kernels with fewer iteration times, reducing network depth and increasing training stability while maintaining the accuracy of the planning process. The gated summarization module enables the network to emphasize the entire planning process, rather than solely relying on the final global planning outcome, by temporally and spatially resampling the entire planning process within the VI module. We conduct experiments on 2D grid world path-finding problems and the Atari Mr. Pac-man environment, demonstrating that GS-VIN outperforms the baseline in terms of single-step accuracy, planning success rate, and overall performance across different map sizes. Additionally, we provide an analysis of the relationship between input size, kernel size, and the number of iterations in VI-based models, which is applicable to a majority of VI-based models and offers valuable insights for researchers and industrial deployment.