Felipe Del Río
Cargo: Asesor de Proyectos Estratégicos
Especialidad:
Investigación en deep learning y generalización de modelos.
Email:
felipe.delrio@cenia.cl
Ingeniero Civil Industrial de la Pontificia Universidad Católica de Chile. Su investigación se centra en el estudio de la generalización en modelos de deep learning, con énfasis en transformers y otros modelos avanzados. Parte relevante de su trabajo aborda el impacto de los datos de entrenamiento en las capacidades de los modelos y en los sesgos de generalización que estos presentan.
PUBLICACIONES
Vision Language Models (VLMs) are designed to extend Large Language Models (LLMs) with visual capabilities, yet in this work we observe a surprising phenomenon: VLMs can outperform their underlying LLMs on purely text-only tasks, particularly in long-context information retrieval. To investigate this effect, we build a controlled synthetic retrieval task and find that a transformer trained only on text achieves perfect in-distribution accuracy but fails to generalize out of distribution, while subsequent training on an image-tokenized version of the same task nearly doubles text-only OOD performance. Mechanistic interpretability reveals that visual training changes the model's internal binding strategy: text-only training encourages positional shortcuts, whereas image-based training disrupts them through spatial translation invariance, forcing the model to adopt a more robust symbolic binding mechanism that persists even after text-only examples are reintroduced. We further characterize how binding strategies vary across training regimes, visual encoders, and initializations, and show that analogous shifts occur during pretrained LLM-to-VLM transitions. Our findings suggest that cross-modal training can enhance reasoning and generalization even for tasks grounded in a single modality.
Edificio de Innovación UC, Piso 2
Vicuña Mackenna 4860
Macul, Chile
Vicuña Mackenna 4860
Macul, Chile