Humans display significant uncertainty when confronted with moral dilemmas, yet the extent of such uncertainty in machines and AI agents remains underexplored. Recent studies have confirmed the overly confident tendencies of machine-generated responses, particularly in large language models (LLMs). As these systems are increasingly embedded in ethical decision-making scenarios, it is important to understand their moral reasoning and the inherent uncertainties in building reliable AI systems. This work examines how uncertainty influences moral decisions in the classical trolley problem, analyzing responses from 32 open-source models and 9 distinct moral dimensions. We first find that variance in model confidence is greater across models than within moral dimensions, suggesting that moral uncertainty is predominantly shaped by model architecture and training method. To quantify uncertainty, we measure binary entropy as a linear combination of total entropy, conditional entropy, and mutual information. To examine its effects, we introduce stochasticity into models via ``dropout'' at inference time. Our findings show that our mechanism increases total entropy, mainly through a rise in mutual information, while conditional entropy remains largely unchanged. Moreover, this mechanism significantly improves human-LLM moral alignment, with correlations in mutual information and alignment score shifts. Our results highlight the potential to better align model-generated decisions and human preferences by deliberately modulating uncertainty and reducing LLMs' confidence in morally complex scenarios.

Large language models leverage both parametric knowledge acquired during pretraining and in-context knowledge provided at inference time.Crucially, when these sources conflict, models arbitrate based on their internal confidence, preferring parametric knowledge for high-confidence facts while deferring to context for less familiar ones.However, the training conditions that give rise to these fundamental behaviors remain unclear.Here we conduct controlled experiments using synthetic corpora to identify the specific data properties that shape knowledge utilization.Our results reveal a counterintuitive finding: the robust, balanced use of both knowledge sources is an emergent property that requires the cooccurrence of three factors typically considered detrimental, including (i) intra-document repetition, (ii) a moderate degree of intra-document inconsistency, and (iii) a skewed knowledge distribution.We further show that these dynamics arise in real-world language model pretraining and analyze how post-training procedures reshape arbitration strategies.Together, our findings provide empirical guidance for designing training data that supports the reliable integration of parametric and in-context knowledge in language models.

Convolutional Neural Networks (CNNs) have profoundly influenced the field of computer vision, drawing significant inspiration from the visual processing mechanisms inherent in the brain. Despite sharing fundamental structural and representational similarities with the biological visual system, differences in local connectivity patterns within CNNs open up an interesting area to explore. In this work, we explore whether integrating biologically observed receptive fields (RFs) can enhance model performance and foster alignment with brain representations. We introduce a novel methodology, termed $L_p$-convolution, which employs the multivariate $L_p$-generalized normal distribution as an adaptable $L_p$-masks, to reconcile disparities between artificial and biological RFs. $L_p$-masks finds the optimal RFs through task-dependent adaptation of conformation such as distortion, scale, and rotation. This allows $L_p$-convolution to excel in tasks that require flexible RF shapes, including not only square-shaped regular RFs but also horizontal and vertical ones. Furthermore, we demonstrate that $L_p$-convolution with biological RFs significantly enhances the performance of large kernel CNNs possibly by introducing structured sparsity inspired by $L_p$-generalized normal distribution in convolution. Lastly, we present that neural representations of CNNs align more closely with the visual cortex when -convolution is close to biological RFs.

Abstract While huge strides have recently been made in language-based machine learning, the ability of artificial systems to comprehend the sequences that comprise animal behavior has been lagging behind. In contrast, humans instinctively recognize behaviors by finding similarities in behavioral sequences. Here, we develop an unsupervised behavior-mapping framework, SUBTLE (spectrogram-UMAP-based temporal-link embedding), to capture comparable behavioral repertoires from 3D action skeletons. To find the best embedding method, we devise a temporal proximity index (TPI) as a new metric to gauge temporal representation in the behavioral embedding space. The method achieves the best TPI score compared to current embedding strategies. Its spectrogram-based UMAP clustering not only identifies subtle inter-group differences but also matches human-annotated labels. SUBTLE framework automates the tasks of both identifying behavioral repertoires like walking, grooming, standing, and rearing, and profiling individual behavior signatures like subtle inter-group differences by age. SUBTLE highlights the importance of temporal representation in the behavioral embedding space for human-like behavioral categorization.

The hippocampus plays a critical role in learning, memory, and spatial representation, processes that depend on the NMDA receptor (NMDAR). Inspired by recent findings that compare deep learning models to the hippocampus, we propose a new nonlinear activation function that mimics NMDAR dynamics. NMDAR-like nonlinearity shifts short-term working memory into long-term reference memory in transformers, thus enhancing a process that is similar to memory consolidation in the mammalian brain. We design a navigation task assessing these two memory functions and show that manipulating the activation function (i.e., mimicking the Mg$^{2+}$-gating of NMDAR) disrupts long-term memory processes. Our experiments suggest that place cell-like functions and reference memory reside in the feed-forward network layer of transformers and that nonlinearity drives these processes. We discuss the role of NMDAR-like nonlinearity in establishing this striking resemblance between transformer architecture and hippocampal spatial representation.