Interpreting Negation in GPT-2: Layer- and Head-Level Causal Analysis
arXiv:2603.12423v1 Announce Type: new Abstract: Negation remains a persistent challenge for modern language models, often causing reversed meanings or factual errors. In this work, we conduct a causal analysis of how GPT-2 Small internally processes such linguistic transformations. We examine...
AgentDrift: Unsafe Recommendation Drift Under Tool Corruption Hidden by Ranking Metrics in LLM Agents
arXiv:2603.12564v1 Announce Type: new Abstract: Tool-augmented LLM agents increasingly serve as multi-turn advisors in high-stakes domains, yet their evaluation relies on ranking-quality metrics that measure what is recommended but not whether it is safe for the user. We introduce a...
LMEB: Long-horizon Memory Embedding Benchmark
arXiv:2603.12572v1 Announce Type: new Abstract: Memory embeddings are crucial for memory-augmented systems, such as OpenClaw, but their evaluation is underexplored in current text embedding benchmarks, which narrowly focus on traditional passage retrieval and fail to assess models' ability to handle...
RTD-Guard: A Black-Box Textual Adversarial Detection Framework via Replacement Token Detection
arXiv:2603.12582v1 Announce Type: new Abstract: Textual adversarial attacks pose a serious security threat to Natural Language Processing (NLP) systems by introducing imperceptible perturbations that mislead deep learning models. While adversarial example detection offers a lightweight alternative to robust training, existing...
98$\times$ Faster LLM Routing Without a Dedicated GPU: Flash Attention, Prompt Compression, and Near-Streaming for the vLLM Semantic Router
arXiv:2603.12646v1 Announce Type: new Abstract: System-level routers that intercept LLM requests for safety classification, domain routing, and PII detection must be both fast and operationally lightweight: they should add minimal latency to every request, yet not require a dedicated GPU...
SectEval: Evaluating the Latent Sectarian Preferences of Large Language Models
arXiv:2603.12768v1 Announce Type: new Abstract: As Large Language Models (LLMs) becomes a popular source for religious knowledge, it is important to know if it treats different groups fairly. This study is the first to measure how LLMs handle the differences...
Adaptive Vision-Language Model Routing for Computer Use Agents
arXiv:2603.12823v1 Announce Type: new Abstract: Computer Use Agents (CUAs) translate natural-language instructions into Graphical User Interface (GUI) actions such as clicks, keystrokes, and scrolls by relying on a Vision-Language Model (VLM) to interpret screenshots and predict grounded tool calls. However,...
Long-form RewardBench: Evaluating Reward Models for Long-form Generation
arXiv:2603.12963v1 Announce Type: new Abstract: The widespread adoption of reinforcement learning-based alignment highlights the growing importance of reward models. Various benchmarks have been built to evaluate reward models in various domains and scenarios. However, a significant gap remains in assessing...
Interpretable Semantic Gradients in SSD: A PCA Sweep Approach and a Case Study on AI Discourse
arXiv:2603.13038v1 Announce Type: new Abstract: Supervised Semantic Differential (SSD) is a mixed quantitative-interpretive method that models how text meaning varies with continuous individual-difference variables by estimating a semantic gradient in an embedding space and interpreting its poles through clustering and...
Multi-Step Semantic Reasoning in Generative Retrieval
arXiv:2603.12368v1 Announce Type: cross Abstract: Generative retrieval (GR) models encode a corpus within model parameters and generate relevant document identifiers directly for a given query. While this paradigm shows promise in retrieval tasks, existing GR models struggle with complex queries...
Reinforcement Learning for Diffusion LLMs with Entropy-Guided Step Selection and Stepwise Advantages
arXiv:2603.12554v1 Announce Type: cross Abstract: Reinforcement learning (RL) has been effective for post-training autoregressive (AR) language models, but extending these methods to diffusion language models (DLMs) is challenging due to intractable sequence-level likelihoods. Existing approaches therefore rely on surrogate likelihoods...
Speech-Worthy Alignment for Japanese SpeechLLMs via Direct Preference Optimization
arXiv:2603.12565v1 Announce Type: cross Abstract: SpeechLLMs typically combine ASR-trained encoders with text-based LLM backbones, leading them to inherit written-style output patterns unsuitable for text-to-speech synthesis. This mismatch is particularly pronounced in Japanese, where spoken and written registers differ substantially in...
Multi-objective Genetic Programming with Multi-view Multi-level Feature for Enhanced Protein Secondary Structure Prediction
arXiv:2603.12293v1 Announce Type: new Abstract: Predicting protein secondary structure is essential for understanding protein function and advancing drug discovery. However, the intricate sequence-structure relationship poses significant challenges for accurate modeling. To address these, we propose MOGP-MMF, a multi-objective genetic programming...
Spatial PDE-aware Selective State-space with Nested Memory for Mobile Traffic Grid Forecasting
arXiv:2603.12353v1 Announce Type: new Abstract: Traffic forecasting in cellular networks is a challenging spatiotemporal prediction problem due to strong temporal dependencies, spatial heterogeneity across cells, and the need for scalability to large network deployments. Traditional cell-specific models incur prohibitive training...
Sinkhorn-Drifting Generative Models
arXiv:2603.12366v1 Announce Type: new Abstract: We establish a theoretical link between the recently proposed "drifting" generative dynamics and gradient flows induced by the Sinkhorn divergence. In a particle discretization, the drift field admits a cross-minus-self decomposition: an attractive term toward...
Bases of Steerable Kernels for Equivariant CNNs: From 2D Rotations to the Lorentz Group
arXiv:2603.12459v1 Announce Type: new Abstract: We present an alternative way of solving the steerable kernel constraint that appears in the design of steerable equivariant convolutional neural networks. We find explicit real and complex bases which are ready to use, for...
Probing Length Generalization in Mamba via Image Reconstruction
arXiv:2603.12499v1 Announce Type: new Abstract: Mamba has attracted widespread interest as a general-purpose sequence model due to its low computational complexity and competitive performance relative to transformers. However, its performance can degrade when inference sequence lengths exceed those seen during...
Embedded Quantum Machine Learning in Embedded Systems: Feasibility, Hybrid Architectures, and Quantum Co-Processors
arXiv:2603.12540v1 Announce Type: new Abstract: Embedded quantum machine learning (EQML) seeks to bring quantum machine learning (QML) capabilities to resource-constrained edge platforms such as IoT nodes, wearables, drones, and cyber-physical controllers. In 2026, EQML is technically feasible only in limited...
Maximizing Incremental Information Entropy for Contrastive Learning
arXiv:2603.12594v1 Announce Type: new Abstract: Contrastive learning has achieved remarkable success in self-supervised representation learning, often guided by information-theoretic objectives such as mutual information maximization. Motivated by the limitations of static augmentations and rigid invariance constraints, we propose IE-CL (Incremental-Entropy...
Swap-guided Preference Learning for Personalized Reinforcement Learning from Human Feedback
arXiv:2603.12595v1 Announce Type: new Abstract: Reinforcement Learning from Human Feedback (RLHF) is a widely used approach to align large-scale AI systems with human values. However, RLHF typically assumes a single, universal reward, which overlooks diverse preferences and limits personalization. Variational...
Human-AI Collaborative Autonomous Experimentation With Proxy Modeling for Comparative Observation
arXiv:2603.12618v1 Announce Type: new Abstract: Optimization for different tasks like material characterization, synthesis, and functional properties for desired applications over multi-dimensional control parameters need a rapid strategic search through active learning such as Bayesian optimization (BO). However, such high-dimensional experimental...
Spend Less, Reason Better: Budget-Aware Value Tree Search for LLM Agents
arXiv:2603.12634v1 Announce Type: new Abstract: Test-time scaling has become a dominant paradigm for improving LLM agent reliability, yet current approaches treat compute as an abundant resource, allowing agents to exhaust token and tool budgets on redundant steps or dead-end trajectories....
LightMoE: Reducing Mixture-of-Experts Redundancy through Expert Replacing
arXiv:2603.12645v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) based Large Language Models (LLMs) have demonstrated impressive performance and computational efficiency. However, their deployment is often constrained by substantial memory demands, primarily due to the need to load numerous expert modules. While...
RetroReasoner: A Reasoning LLM for Strategic Retrosynthesis Prediction
arXiv:2603.12666v1 Announce Type: new Abstract: Retrosynthesis prediction is a core task in organic synthesis that aims to predict reactants for a given product molecule. Traditionally, chemists select a plausible bond disconnection and derive corresponding reactants, which is time-consuming and requires...
Disentangled Latent Dynamics Manifold Fusion for Solving Parameterized PDEs
arXiv:2603.12676v1 Announce Type: new Abstract: Generalizing neural surrogate models across different PDE parameters remains difficult because changes in PDE coefficients often make learning harder and optimization less stable. The problem becomes even more severe when the model must also predict...
Before quantum computing arrives, this startup wants enterprises already running on it
After selling his AI startup to AMD for $665 million, Peter Sarlin is back with Qutwo, a new venture building the infrastructure it believes enterprises will need when quantum computing finally arrives.
PACED: Distillation at the Frontier of Student Competence
arXiv:2603.11178v1 Announce Type: new Abstract: Standard LLM distillation wastes compute on two fronts: problems the student has already mastered (near-zero gradients) and problems far beyond its reach (incoherent gradients that erode existing capabilities). We show that this waste is not...
Summarize Before You Speak with ARACH: A Training-Free Inference-Time Plug-In for Enhancing LLMs via Global Attention Reallocation
arXiv:2603.11067v1 Announce Type: new Abstract: Large language models (LLMs) achieve remarkable performance, yet further gains often require costly training. This has motivated growing interest in post-training techniques-especially training-free approaches that improve models at inference time without updating weights. Most training-free...
MDER-DR: Multi-Hop Question Answering with Entity-Centric Summaries
arXiv:2603.11223v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) over Knowledge Graphs (KGs) suffers from the fact that indexing approaches may lose important contextual nuance when text is reduced to triples, thereby degrading performance in downstream Question-Answering (QA) tasks, particularly for...
Algorithmic Consequences of Particle Filters for Sentence Processing: Amplified Garden-Paths and Digging-In Effects
arXiv:2603.11412v1 Announce Type: new Abstract: Under surprisal theory, linguistic representations affect processing difficulty only through the bottleneck of surprisal. Our best estimates of surprisal come from large language models, which have no explicit representation of structural ambiguity. While LLM surprisal...