TrueBrief: Faithful Summarization through Small Language Models
This work addresses the challenge of reliable text summarization for security-critical applications, though it appears incremental as it builds on existing preference optimization methods.
The paper tackles the problem of hallucinations in large language models for security-critical domains by introducing TrueBrief, a framework that enhances faithfulness in small language models for text summarization through preference optimization with controlled hallucination injection, achieving improved performance as demonstrated in their experiments.
Large language models (LLMs) have exhibited remarkable proficiency in generating high-quality text; however, their propensity for producing hallucinations poses a significant challenge for their deployment in security-critical domains. In this work, we present TrueBrief, an end-to-end framework specifically designed to enhance the faithfulness of small LLMs (SLMs) primarily for the task of text summarization through a preference-optimization paradigm. Central to our framework is a data generation module that facilitates controlled hallucination injection to generate synthetic preference data. Our work provides insights into the impact of data quality and model size on preference-based optimization, highlighting the conditions under which these methods are most effective.