Can LLMs Track Their Output Length? A Dynamic Feedback Mechanism for Precise Length Regulation

TL;DR

This paper introduces a dynamic feedback mechanism for LLMs that improves their ability to generate text of precise lengths, addressing a key challenge in controlled text generation.

Contribution

The work presents a novel, training-free length regulation method with dynamic feedback, enhancing LLMs' ability to meet specific length constraints across tasks.

Findings

Significantly improves length adherence in summarization and biography tasks

Enables adaptive adjustments during generation without retraining

Generalizes well to various text-generation tasks after fine-tuning

Abstract

Precisely controlling the length of generated text is a common requirement in real-world applications. However, despite significant advancements in following human instructions, Large Language Models (LLMs) still struggle with this task. In this work, we demonstrate that LLMs often fail to accurately measure their response lengths, leading to poor adherence to length constraints. To address this issue, we propose a novel length regulation approach that incorporates dynamic length feedback during generation, enabling adaptive adjustments to meet target lengths. Experiments on summarization and biography tasks show our training-free approach significantly improves precision in achieving target token, word, or sentence counts without compromising quality. Additionally, we demonstrate that further supervised fine-tuning allows our method to generalize effectively to broader text-generation tasks.