TrueBrief: Faithful Summarization through Small Language Models

TL;DR

TrueBrief introduces a framework that improves the faithfulness of small language models in text summarization by using preference optimization and synthetic data generation to reduce hallucinations.

Contribution

The paper presents a novel end-to-end approach for enhancing faithfulness in small language models through controlled data generation and preference-based optimization.

Findings

Data quality significantly impacts optimization effectiveness

Model size influences the success of preference-based methods

Controlled hallucination injection improves summarization faithfulness

Abstract

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.