arXiv:2505.13353v4 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly deployed for understanding large codebases, but whether they understand operational semantics of long code context or rely on pattern matching shortcuts remains unclear. We distinguish between lexical recall (retrieving…
arXiv:2604.16411v1 Announce Type: new Abstract: We study asynchronous alignment, a first-class multimodal learning setting in which a dense primary stream must be fused with sporadic external context whose value depends on when it arrives. Unlike standard multimodal benchmarks that assume…
arXiv:2601.13707v2 Announce Type: replace-cross Abstract: Hallucinations in large vision–language models (LVLMs) often arise when language priors dominate over visual evidence, leading to object misidentification and visually inconsistent descriptions. We address this problem by framing hallucination mitigation as contrastive guidance that…
arXiv:2604.16423v1 Announce Type: new Abstract: Defensive training methods such as positive preventative steering (PPS) and inoculation prompting (IP) offer surprising results through seemingly similar processes: both add trait-inducing objects to large language models (LLMs) during training, and both defend the…
arXiv:2604.18137v1 Announce Type: cross Abstract: Processing-in-Memory (PIM) architectures offer a promising solution to the memory bottlenecks in data-intensive machine learning, yet often overlook the growing challenge of activation memory footprint. Conventional PIM approaches struggle with massive KV cache sizes generated…
arXiv:2604.16426v1 Announce Type: new Abstract: As modern deep learning architectures grow in complexity, representational ambiguity emerges as a critical barrier to their interpretability and reliable merging. For ReLU networks, identical functional mappings can be achieved through entirely different weight configurations…
arXiv:2604.18566v1 Announce Type: cross Abstract: We present a systematic evaluation of large language model families — spanning both proprietary cloud APIs and locally-hosted open-source models — on two purpose-built benchmarks for System Dynamics AI assistance: the textbf{CLD Leaderboard} (53 tests,…
arXiv:2604.16428v1 Announce Type: new Abstract: Time series foundation models (TSFMs) are widely used as generic feature extractors, yet the notion of non-stationarity in their embedding spaces remains poorly understood. Recent work often conflates non-stationarity with distribution shift, blurring distinctions fundamental…
arXiv:2604.16358v1 Announce Type: new Abstract: MLLMs are increasingly deployed in multi-turn settings, where attackers can escalate unsafe intent through the evolving visual-text history and exploit long-context safety decay. Yet safety alignment is still dominated by single-turn data and fixed-template dialogues,…
arXiv:2506.12622v2 Announce Type: replace Abstract: Deep reinforcement learning (RL) has achieved remarkable success, yet its deployment in real-world scenarios is often limited by vulnerability to environmental uncertainties. Distributionally robust RL (DR-RL) algorithms have been proposed to resolve this challenge, but…
