LATIDIA · Investigación

Grandes modelos lingüísticos (LLM)

arXiv: 2610.38120v1Tipo de anuncio: nuevo Resumen: los modelos de lenguaje grandes (LLM) tratan todos los tokens de contexto de manera uniforme, sin separación explícita entre las instrucciones ejecutables y los datos pasivos. Esto conduce a una inyección rápida

La noticia

arXiv:2610.38120v1 Announce Type: new Abstract: Large Language Models (LLMs) treat all context tokens uniformly, lacking explicit separation between executable instructions and passive data. This leads to prompt injection vulnerabilities, ambiguous instruction binding, and unreliable execution control. We present a post-training methodology using LLMON (LLM Object Notation), a framework that enables separating instructions from data and requires explicit identifier-based execution. We construct a large-scale structured training corpus of 4.6B tokens spanning general, RAG (Retrieval-Augmented Generation), and tool-calling data, augmented with distractor-infused variants in which competing instructions and irrelevant data segments are present in context but not referenced. Post-training two Granite-4.1 base models (3B and 8B) through supervised fine-tuning followed by reinforcement