LATIDIA · Robótica

Los caracterizadores multidimensionales rinden dirección de grano fino y conformación de datos en modelos de cimentación de robots

arXiv:2610.20767v1 Anuncio Tipo: nuevo Resumen: ¿Puede un mejor descubrimiento de características mejorar el control sobre el comportamiento del robot a través de intervenciones internas y datos de entrenamiento? Trabajos recientes en modelos de visión y lenguaje revelan múltiples

La noticia

arXiv:2610.20767v1 Announce Type: new Abstract: Can better feature discovery improve control over robot behavior through internal interventions and training data? Recent work in vision and language models reveals multidimensional features, motivating block-sparse featurizers (BSFs), which represent each feature in a low-dimensional subspace. We investigate whether recovering this structure provides an interpretable interface for steering vision-language-action models and shaping the data from which they learn. We show that BSFs recover several multi-dimensional task-relevant features underlying the model's computation (e.g., perception, language, proprioceptive state, episode phase, and control variables). We causally validate these features through fine-grained steering: targeted changes in visual perception, semantic representation, and spatial localization produce predictable changes in downstream