Inverse Language Modeling towards Robust and Grounded LLMs

Davide Gabrielli, Simone Sestito, Iacopo Masi

Keywords: Large Language Models, LLMs, Robustness, Adversarial Attacks, GCG Attack, Interpretability, Gradient-based Inversion, Grounded LLMs, Red Teaming, LLM Inversion

Abstract: Interpretability and robustness remain major challenges for modern Large Language Models (LLMs), especially in settings where conventional evaluation or auditing tools are limited. To address this, we propose Inverse Language Modeling (ILM), a unified training framework that jointly enhances robustness to adversarial perturbations and enables a novel form of gradient-based interpretability. Rather than reconstructing exact input prompts, ILM encourages LLMs to develop gradient-aligned internal representations that allow the model to approximate plausible input patterns underlying a given output. This approximate inversion provides a new mechanism for analyzing model behavior, identifying potential triggers for unsafe generations, and providing a diagnostic signal that may support future auditing workflows. Our results show that ILM can simultaneously improve robustness and produce meaningful inversion signals, laying a foundation for LLMs that are not only more resilient but also more transparent and analyzable.

Submission Number: 17