DPWriter: Reinforcement Learning with Diverse Planning Branching for Creative Writing

TL;DR

This paper introduces DPWriter, an RL framework for creative writing that enhances output diversity by explicitly planning intermediate steps and encouraging diverse trajectories, outperforming existing methods.

Contribution

It proposes a novel RL approach with diverse planning branching and a group-aware diversity reward to improve output diversity in creative writing tasks.

Findings

Significantly improves output diversity in creative writing benchmarks.

Maintains high quality of generated content.

Outperforms existing baseline methods.

Abstract

Reinforcement learning (RL)-based enhancement of large language models (LLMs) often leads to reduced output diversity, undermining their utility in open-ended tasks like creative writing. Current methods lack explicit mechanisms for guiding diverse exploration and instead prioritize optimization efficiency and performance over diversity. This paper proposes an RL framework structured around a semi-structured long Chain-of-Thought (CoT), in which the generation process is decomposed into explicitly planned intermediate steps. We introduce a Diverse Planning Branching method that strategically introduces divergence at the planning phase based on diversity variation, alongside a group-aware diversity reward to encourage distinct trajectories. Experimental results on creative writing benchmarks demonstrate that our approach significantly improves output diversity without compromising generation quality, consistently outperforming existing baselines.