Breaking the Trade-Off Between Faithfulness and Expressiveness for Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yang, Chenxu, Si, Qingyi, Lin, Zheng
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912555001184256
author Yang, Chenxu
Si, Qingyi
Lin, Zheng
author_facet Yang, Chenxu
Si, Qingyi
Lin, Zheng
contents Grounding responses in external knowledge represents an effective strategy for mitigating hallucinations in Large Language Models (LLMs). However, current LLMs struggle to seamlessly integrate knowledge while simultaneously maintaining faithfulness (or fidelity) and expressiveness, capabilities that humans naturally possess. This limitation results in outputs that either lack support from external knowledge, thereby compromising faithfulness, or appear overly verbose and unnatural, thus sacrificing expressiveness. In this work, to break the trade-off between faithfulness and expressiveness, we propose Collaborative Decoding (CoDe), a novel approach that dynamically integrates output probabilities generated with and without external knowledge. This integration is guided by distribution divergence and model confidence, enabling the selective activation of relevant and reliable expressions from the model's internal parameters. Furthermore, we introduce a knowledge-aware reranking mechanism that prevents over-reliance on prior parametric knowledge while ensuring proper utilization of provided external information. Through comprehensive experiments, our plug-and-play CoDe framework demonstrates superior performance in enhancing faithfulness without compromising expressiveness across diverse LLMs and evaluation metrics, validating both its effectiveness and generalizability.
format Preprint
id arxiv_https___arxiv_org_abs_2508_18651
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Breaking the Trade-Off Between Faithfulness and Expressiveness for Large Language Models
Yang, Chenxu
Si, Qingyi
Lin, Zheng
Computation and Language
Artificial Intelligence
Grounding responses in external knowledge represents an effective strategy for mitigating hallucinations in Large Language Models (LLMs). However, current LLMs struggle to seamlessly integrate knowledge while simultaneously maintaining faithfulness (or fidelity) and expressiveness, capabilities that humans naturally possess. This limitation results in outputs that either lack support from external knowledge, thereby compromising faithfulness, or appear overly verbose and unnatural, thus sacrificing expressiveness. In this work, to break the trade-off between faithfulness and expressiveness, we propose Collaborative Decoding (CoDe), a novel approach that dynamically integrates output probabilities generated with and without external knowledge. This integration is guided by distribution divergence and model confidence, enabling the selective activation of relevant and reliable expressions from the model's internal parameters. Furthermore, we introduce a knowledge-aware reranking mechanism that prevents over-reliance on prior parametric knowledge while ensuring proper utilization of provided external information. Through comprehensive experiments, our plug-and-play CoDe framework demonstrates superior performance in enhancing faithfulness without compromising expressiveness across diverse LLMs and evaluation metrics, validating both its effectiveness and generalizability.
title Breaking the Trade-Off Between Faithfulness and Expressiveness for Large Language Models
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2508.18651