arXiv:2605.00972v1 Announce Type: cross Abstract: Earth system science is producing increasingly large, high-dimensional datasets from physics based Earth system models to AI-based weather and climate models. Embedding-based representations can make these data searchable through similarity search and analog retrieval, but nearest neighbors in latent space are not automatically scientifically meaningful: it may reflect real weather […]
LLM Ghostbusters: Surgical Hallucination Suppression via Adaptive Unlearning
arXiv:2605.01047v1 Announce Type: cross Abstract: Hallucinations, outputs that sound plausible but are factually incorrect, remain an open challenge for deployed LLMs. In code generation, models frequently hallucinate non-existent software packages, recommending imports and installation commands for fictional libraries. This creates a critical supply-chain vulnerability: an attacker can proactively register such packages on public registries with […]
When Less is Enough: Efficient Inference via Collaborative Reasoning
arXiv:2605.01111v1 Announce Type: cross Abstract: In this work, we introduce DUET (Dual-model Efficient Two-stage inference), a collaborative inference framework in which a capable model and a lightweight model work together to solve a task. Relying on a single large model to perform end-to-end reasoning and prediction often incurs substantial inference cost. In contrast, DUET decomposes […]
TetraJet-v2: Accurate NVFP4 Training for Large Language Models with Oscillation Suppression and Outlier Control
arXiv:2510.27527v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) training is prohibitively expensive, driving interest in low-precision fully-quantized training (FQT). While novel 4-bit formats like NVFP4 offer substantial efficiency gains, achieving near-lossless training at such low precision remains challenging. We introduce TetraJet-v2, an end-to-end 4-bit FQT method that leverages NVFP4 for activations, weights, and gradients […]
MAD-OPD: Breaking the Ceiling in On-Policy Distillation via Multi-Agent Debate
arXiv:2605.01347v1 Announce Type: cross Abstract: On-policy distillation (OPD) trains a student on its own trajectories under token-level teacher supervision, but existing methods are capped by a single-teacher capability ceiling: when the teacher errs, the student inherits the error. OPD also remains largely unexplored in agentic tasks, where per-step errors compound across long trajectories and destabilize […]
EmoMM: Benchmarking and Steering MLLM for Multimodal Emotion Recognition under Conflict and Missingness
arXiv:2605.01024v1 Announce Type: cross Abstract: Multimodal Emotion Recognition (MER) is critical for interpreting real-world interactions. While Multimodal Large Language Models (MLLM) have shown promise in MER, their internal decision-making mechanisms under modality conflict and missingness remain largely underexplored. In this paper, to systematically investigate these behaviors, we introduce EmoMM, a comprehensive benchmark featuring modality-aligned, conflict, […]
Minimizing Collateral Damage in Activation Steering
arXiv:2605.01167v1 Announce Type: cross Abstract: Activation steering is a method for controlling Large Language Model (LLM) behavior by intervening in its internal representations to increase the alignment with a specific target feature direction. However, standard interventions, such as vector addition, often cause “collateral damage”, defined as unintended changes in the alignment of activations along other […]
When Less Is More: Simplicity Beats Complexity for Physics-Constrained InSAR Phase Unwrapping
arXiv:2605.00896v1 Announce Type: cross Abstract: Operational phase unwrapping is the primary computational bottleneck in InSAR-based volcanic and seismic monitoring. We challenge the industry trend of adopting high-complexity computer vision architectures, such as attention mechanisms, without validating their suitability for physics-constrained geophysical regression. We present the first large-scale architectural ablation study on a global LiCSAR benchmark […]
Voice Mapping of Text-to-Speech Systems: A Metric-Based Approach for Voice Quality Assessment
arXiv:2605.00861v1 Announce Type: cross Abstract: This study investigates voice mapping as an evaluation framework for text-to-speech (TTS) synthesis quality. The study analyzes six TTS models, including historical and recent ones. The metrics are crest factor, spectrum balance, and cepstral peak prominence (CPPs). We investigated 6 influential TTS models: Merlin, Tacotron 2, Transformer TTS, FastSpeech 2, […]
VisInject: Disruption != Injection — A Dual-Dimension Evaluation of Universal Adversarial Attacks on Vision-Language Models
arXiv:2605.01449v1 Announce Type: cross Abstract: Universal adversarial attacks on aligned multimodal large language models are increasingly reported with attack success rates in the 60-80% range, suggesting the visual modality is highly vulnerable to imperceptible perturbations as a prompt-injection channel. We argue that this number conflates two distinct events: (i) the model’s output was perturbed (Influence), […]
Where Do Prompt Perturbations Break Generation? A Segment-Level View of Robustness in LoRA-Tuned Language Models
arXiv:2605.01605v1 Announce Type: cross Abstract: Large language models are sensitive to minor prompt perturbations, yet existing robustness methods usually enforce consistency at the whole-sequence level. This holistic view can hide an important failure mode: a perturbed response may remain globally similar to the clean one while drifting on a critical entity, relation, or conclusion. We […]
GeoSAE: Geometric Prior-Guided Layer-Wise Sparse Autoencoder Annotation of Brain MRI Foundation Models
arXiv:2605.01829v1 Announce Type: cross Abstract: Brain MRI foundation models learn rich representations of anatomy, but interpreting what clinical information they encode remains an open problem. Standard sparse autoencoders (SAEs) suffer from severe feature collapse in deep transformer layers, and in Alzheimer’s disease (AD) research, aging confounds nearly every clinical variable, making naive annotation unreliable. We […]