arXiv:2506.10138v3 Announce Type: replace-cross Abstract: We partially reverse-engineer a convolutional recurrent neural network (RNN) trained with model-free reinforcement learning to play the box-pushing game Sokoban. We find that the RNN stores future moves (plans) as activations in particular channels of the hidden state, which we call path channels. A high activation in a particular location […]
Detection Without Correction: A Two-Parameter Decomposition of Multi-Stage LLM Pipelines
arXiv:2605.27559v1 Announce Type: cross Abstract: Multi-stage LLM pipelines that perform multi-agent debate, intrinsic self-correction, or retrieval-augmented verification exhibit puzzling aggregate behaviors: accuracy plateaus and reversals across rounds, non-replication of debate gains on contemporary frontier models, intrinsic self-correction degradation, and qualitative cross-provider divergence in debate dynamics. Downstream agent response can be operationalized as two coupled decisions: […]
Laguna M.1/XS.2 Technical Report
arXiv:2605.27605v1 Announce Type: new Abstract: We present Laguna M.1 and Laguna XS.2, two Mixture-of-Experts foundation models built for long-horizon, agentic coding: M.1 has $225.8$B total parameters ($23.4$B activated per token) and XS.2 has $33.4$B total ($3$B activated). Both models were trained from scratch end-to-end inside the same internal system that we refer to as our […]
Supervised Distributional Reduction via Optimal Transport and Dependence Maximization
arXiv:2605.27619v1 Announce Type: cross Abstract: Learning representations that capture both intrinsic data geometry and target-relevant structure remains a fundamental challenge, particularly in settings where data reduction must balance compression with predictive fidelity. While distributional reduction-encompassing joint clustering and dimensionality reduction-offers a principled way to summarize data, its supervised variants remain relatively under-explored, despite the importance […]
NCSAM Noise-Compensated Sharpness-Aware Minimization for Noisy Label Learning
arXiv:2601.19947v2 Announce Type: replace-cross Abstract: Learning from Noisy Labels (LNL) remains a fundamental challenge in deep learning because real-world datasets often contain corrupted annotations. Most existing methods rely on label correction or sample selection mechanisms. In contrast, we study LNL from an optimization perspective by establishing a theoretical connection between label noise and the flatness-seeking […]
Backdoor Attacks on Fault Detection and Localization in Cyber-Physical Systems
arXiv:2605.27674v1 Announce Type: cross Abstract: Cyber-Physical Systems (CPS) integrate sensing, communication, computation, and control to support critical infrastructure, including smart grids, industrial automation, and control systems. In the electrical utility domain, various controllers are used in CPS to ensure the system detects and recovers from faults, such as voltage fluctuations, and to perform load balancing […]
Reasoning and Planning with Dynamically Changing Norms
arXiv:2605.27622v1 Announce Type: new Abstract: To safely interact with humans, AI agents must both know our norms and consider them during planning. However, such norm-guided planning has been less explored, only within communities of artificial agents, and has ignored the dynamic nature of norms. This paper instead presents an approach to guiding planning with dynamically […]
Reading or Guessing? Visual Grounding Failures of Vision-Language Models for OCR in Ancient Greek Editions
arXiv:2605.27750v1 Announce Type: cross Abstract: Recent work has shown that Vision-Language Models (VLMs) used for optical character recognition (OCR) can generate plausible but visually unsupported text, suggesting reliance on language priors. Comparing open-weight VLMs with traditional OCR baselines on low-resource Ancient Greek critical editions, we show that VLM errors often remain fluent even when wrong, […]
COTTA: Context-Aware Transfer Adaptation for Trajectory Prediction in Autonomous Driving
arXiv:2604.00402v2 Announce Type: replace-cross Abstract: Developing robust models to accurately predict the trajectories of surrounding agents is fundamental to autonomous driving safety. However, most public datasets, such as the Waymo Open Motion Dataset and Argoverse, are collected in Western road environments and do not reflect the unique traffic patterns, infrastructure, and driving behaviors of other […]
Residualized Temporal Sparse Autoencoders for Interpreting Diffusion Models
arXiv:2605.27813v1 Announce Type: cross Abstract: Text-to-image diffusion models generate images through an iterative denoising process, so internal neural layers produce trajectories of activations rather than single static representations. Sparse autoencoders (SAEs) have recently been used to decompose diffusion activations into interpretable feature directions, but most approaches analyze activations at individual timesteps or condition on time […]
Fine-Tuned LLM as a Complementary Predictor Improving Ads System
arXiv:2605.27856v1 Announce Type: cross Abstract: Recommendation systems power engagement and monetization across feeds, ads, and short-video platforms, but translating the latest advances in Large Language Models into Recommendation Systems (RecSys) gains remains rare, particularly in advertising and production-scale real-world industry setups. Prior real-world LLM successes typically fall into three buckets: (a) generative retrieval that directly […]
One LR Doesn’t Fit All: Heavy-Tail Guided Layerwise Learning Rates for LLMs
arXiv:2605.22297v3 Announce Type: replace-cross Abstract: Learning rate configuration is a fundamental aspect of modern deep learning. The prevailing practice of applying a uniform learning rate across all layers overlooks the structural heterogeneity of Transformers, potentially limiting their effectiveness as the backbone of Large Language Models (LLMs). In this paper, we introduce Layerwise Learning Rate (LLR), […]