The evolution of cloud datacenter infrastructure, driven by ever-increasing bandwidth demands, has shifted performance bottlenecks from network hardware to server-side protocol processing. To mitigate per-CPU processing bottlenecks, load-balancing schemes such as Receive Side Scaling (RSS) and accelerated Receive Flow Steering (aRFS) have been introduced to distribute network flows across multiple CPU cores for incoming packet processing. While improving overall core utilization and cache locality, these schemes do not consider recent advancements such as non-uniform memory access (NUMA)-based architectures and direct cache access (DCA), leading to suboptimal performance, particularly when large and small flows are mixed. In this paper, we introduce a new receive flow steering scheme, mixed Flow Steering (mFS), which disaggregates large and small flows to optimize network flow steering in NUMA-based architectures. Our approach incorporates a flow monitoring module to classify flows based on cumulative data volume, a flow core migration mechanism that aligns application processing with the appropriate NUMA node, and adaptive handling of multi-large flow contention. Our experimental evaluations demonstrate that, by leveraging DCA, mFS significantly improves total throughput for large flows by 42.12% compared to existing schemes while maintaining comparable throughput for small flows across various mixed traffic scenarios and real-world applications.

Recent remarkable advances in Large Language Models (LLMs) have led to innovations in various domains such as education, healthcare, and finance, while also raising serious concerns that they can be easily misused for malicious purposes. Most previous research has focused primarily on observing how jailbreak attack techniques bypass safety mechanisms like Reinforcement Learning through Human Feedback (RLHF). However, whether there are neurons within LLMs that directly govern aggression has not been sufficiently investigated. To fill this gap, this study identifies specific neurons (“aggression neurons”) closely related to the expression of aggression and systematically analyzes how manipulating them affects the model’s overall aggression. Specifically, using a large-scale synthetic text corpus (aggressive and non-aggressive), we measure the activation frequency of each neuron, then apply masking and activation techniques to quantitatively evaluate changes in aggression by layer and by manipulation ratio. Experimental results show that, in all models, manipulating only a small number of neurons can increase aggression by up to 33%, and the effect is even more extreme when aggression neurons are concentrated in certain layers. Moreover, even models of the same scale exhibit nonlinear changes in aggression patterns, suggesting that simple external safety measures alone may not be sufficient for complete defense.

Instruction-tuning enhances the ability of large language models (LLMs) to follow user instructions more accurately, improving usability while reducing harmful outputs.However, this process may increase the model's dependence on user input, potentially leading to the unfiltered acceptance of misinformation and the generation of hallucinations.Existing studies primarily highlight that LLMs are receptive to external information that contradict their parametric knowledge, but little research has been conducted on the direct impact of instructiontuning on this phenomenon.In our study, we investigate the impact of instruction-tuning on LLM's susceptibility to misinformation.Our analysis reveals that instruction-tuned LLMs are significantly more likely to accept misinformation when it is presented by the user.A comparison with base models shows that instruction-tuning increases reliance on userprovided information, shifting susceptibility from the assistant role to the user role.Furthermore, we explore additional factors influencing misinformation susceptibility, such as the role of the user in prompt structure, misinformation length, and the presence of warnings in the system prompt.Our findings underscore the need for systematic approaches to mitigate unintended consequences of instruction-tuning and enhance the reliability of LLMs in realworld applications.