Related papers: SynthSAEBench: Evaluating Sparse Autoencoders on S…

Large language models (LLMs) have demonstrated remarkable performance, yet their diverse strengths and weaknesses prevent any single LLM from achieving dominance across all tasks. Ensembling multiple LLMs is a promising approach to generate…

Large Audio Language Models (LALMs) have been widely applied in real-time scenarios, such as in-car assistants and online meeting comprehension. In practice, audio inputs are often corrupted by device and environmental noise, leading to…

Fine-grained steering of language model outputs is essential for safety and reliability. Prompting and finetuning are widely used to achieve these goals, but interpretability researchers have proposed a variety of representation-based…

We describe HypotheSAEs, a general method to hypothesize interpretable relationships between text data (e.g., headlines) and a target variable (e.g., clicks). HypotheSAEs has three steps: (1) train a sparse autoencoder on text embeddings to…

Spoken Language Models (SLMs) have emerged as a promising paradigm for speech synthesis by bypassing explicit grapheme-to-phoneme pipelines. However, their effectiveness in low-resource languages remains fundamentally limited by the…

Training large language models (LLMs) from scratch requires significant computational resources, driving interest in developing smaller, domain-specific LLMs that maintain both efficiency and strong task performance. Medium-sized models…

This paper introduces Synthetic Enclosed Echoes (SEE), a novel dataset designed to enhance robot perception and 3D reconstruction capabilities in underwater environments. SEE comprises high-fidelity synthetic sonar data, complemented by a…

SpoofTrackBench is a reproducible, modular benchmark for evaluating adversarial robustness in real-time localization and tracking (RTLS) systems under radar spoofing. Leveraging the Hampton University Skyler Radar Sensor dataset, we…

Context engineering has emerged as a pivotal paradigm for unlocking the potential of Large Language Models (LLMs) in Software Engineering (SE) tasks, enabling performance gains at test time without model fine-tuning. Despite its success,…

Sparse autoencoders (SAEs) are useful for detecting and steering interpretable features in neural networks, with particular potential for understanding complex multimodal representations. Given their ability to uncover interpretable…

Stochastic Partial Differential Equations (SPDEs) driven by random noise play a central role in modeling physical processes with rough spatio-temporal dynamics, such as turbulence flows, superconductors, and quantum dynamics. Although…

We propose a method for using synthetic data to help learning classifiers. Synthetic data, even is generated based on real data, normally results in a shift from the distribution of real data in feature space. To bridge the gap between the…

Radiological services are experiencing unprecedented demand, leading to increased interest in automating radiology report generation. Existing Vision-Language Models (VLMs) suffer from hallucinations, lack interpretability, and require…

Intermediate layers of large language models (LLMs) best predict human brain responses to language, one of the most robust findings in computational neurolinguistics, yet why remains mechanistically unexplained. We address this gap by…

Sparse Autoencoders (SAEs) have recently emerged as powerful tools for interpreting and steering the internal representations of large language models (LLMs). However, conventional approaches to analyzing SAEs typically rely solely on…

Automated Synthesis Planning has recently re-emerged as a research area at the intersection of chemistry and machine learning. Despite the appearance of steady progress, we argue that imperfect benchmarks and inconsistent comparisons mask…

The ability of large language models (LLMs) to follow instructions is crucial for their practical applications, yet the underlying mechanisms remain poorly understood. This paper presents a novel framework that leverages sparse autoencoders…

Motivated by the hypothesis that neural network representations encode abstract, interpretable features as linearly accessible, approximately orthogonal directions, sparse autoencoders (SAEs) have become a popular tool in interpretability.…

Recent progress has expanded the use of large language models (LLMs) in drug discovery, including synthesis planning. However, objective evaluation of retrosynthesis performance remains limited. Existing benchmarks and metrics typically…

Understanding and mitigating the potential risks associated with foundation models (FMs) hinges on developing effective interpretability methods. Sparse Autoencoders (SAEs) have emerged as a promising tool for disentangling FM…