SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation

2025-02-19

AI Summary

Large language models (LLMs) frequently exhibit hallucination—generating unfaithful outputs inconsistent with input contexts—in conditional text generation tasks such as summarization and data-to-text generation, primarily due to overreliance on spurious statistical patterns in training data. To address this, we propose the first fully self-supervised framework that requires no human annotations: it automatically constructs unfaithful samples via controllable perturbations of input contexts, then jointly optimizes faithfulness through preference learning and contrastive fine-tuning. We further introduce a dual evaluation protocol integrating LLM-based automatic assessment and human evaluation for robust fidelity validation. Experiments demonstrate that our method significantly outperforms existing self-supervised approaches across multiple faithfulness metrics, substantially improving output fidelity in both summarization and data-to-text generation. This work establishes a novel paradigm for enhancing faithfulness in low-resource settings without manual supervision.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Diffusion Models for Vision

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Large Language Models (LLMs), when used for conditional text generation, often produce hallucinations, i.e., information that is unfaithful or not grounded in the input context. This issue arises in typical conditional text generation tasks, such as text summarization and data-to-text generation, where the goal is to produce fluent text based on contextual input. When fine-tuned on specific domains, LLMs struggle to provide faithful answers to a given context, often adding information or generating errors. One underlying cause of this issue is that LLMs rely on statistical patterns learned from their training data. This reliance can interfere with the model's ability to stay faithful to a provided context, leading to the generation of ungrounded information. We build upon this observation and introduce a novel self-supervised method for generating a training set of unfaithful samples. We then refine the model using a training process that encourages the generation of grounded outputs over unfaithful ones, drawing on preference-based training. Our approach leads to significantly more grounded text generation, outperforming existing self-supervised techniques in faithfulness, as evaluated through automatic metrics, LLM-based assessments, and human evaluations.

Problem

• Addressing hallucinations in text generation
• Improving faithfulness in conditional text generation
• Reducing ungrounded information in LLM outputs

Innovation

• self-supervised framework
• unfaithful samples training
• preference-based training refinement