PaperlayerView exact revision on arXiv

Source-linked AI summary

Scaling Creative Writing Beyond Story-Centric Data with Attribute-Guided Genre Expansion

Hwan Chang, Yongil Kim, Heuiyeen Yeen, Yireun Kim, Jinsik Lee, Hwanhee Lee

arXiv:2608.13947v1cs.CL

TL;DR

Story-centric creative-writing data does not adequately represent the structural conventions of diverse formats. This paper introduces attribute-guided genre expansion and finds that its 50K-example, 13-genre dataset improves creative-writing performance and output novelty.

  • Problem

    Existing creative-writing resources are heavily story-centric, despite diverse formats having distinct structural, stylistic, functional, and formatting conventions.

    Source: §1 · pp. 1–2

  • Method

    The framework expands human-authored story prompts into genre-faithful instruction data using manually curated attributes that control genre-specific conventions.

    Source: §2 · p. 3

  • Results

    Models fine-tuned on the Multi-Genre Collection outperform base models, writing-specialized baselines, and existing writing-dataset models across benchmarks, while novelty rises with genre coverage.

    Source: §1 · p. 2, §3.2 · p. 4

  • Takeaways & Limitations

    Controlled expansion across 13 genres provides more transferable creative-writing training data and enhances output novelty.

    Source: §3.2 · p. 4, §5 · p. 5

Abstract

from arXiv · show

High-quality creative writing data for large language models (LLMs) remains dominated by story-centric data, limiting models' ability to follow the structural and functional conventions of diverse creative formats. We propose an attribute-guided genre expansion framework for scaling creative writing data beyond story generation. By separating thematic breadth from genre-form control, our framework leverages human-authored story prompts as diverse creative seeds, while utilizing manually curated genre attributes to enforce distinct structural, stylistic, and formatting conventions. We combine these to prompt strong LLMs for genre-faithful query-response pairs, which are then quality-filtered. Applying this framework, we construct the Multi-Genre Collection, a 50K-example corpus spanning 13 creative genres, including story, rap, lyrics, scripts, game design, character design, and other creative formats. Experiments across out-of-distribution writing benchmarks and held-out genre diagnostics demonstrate that models fine-tuned on our data consistently surpass not only base models and writing-specialized baselines, but also models trained on existing writing corpora. Genre-count ablations further indicate that controlled genre expansion, rather than story-centric scaling alone, is a key driver of robust creative writing capability.

1 Introduction

Creative-writing resources remain heavily story-centric, leaving non-story genres underrepresented and their structural conventions insufficiently encoded. The paper addresses this gap with attribute-guided genre expansion, a 50K-instance corpus spanning 13 genres that improves creative-writing capability across benchmarks.

Motivation

  • Public creative-writing resources are dominated by story-centric narrative generation, despite users requesting formats such as rap verses and game design documents.These genres differ beyond topic, requiring distinct formal conventions.Source: §1 · p. 1
  • Scaling story-centric data alone does not systematically encode the formal constraints required by non-story genres.Generic synthesis can produce topically diverse but formally under-specified prompts, motivating human guidance for genre-form constraints.Source: §1 · pp. 1–2

Approach

  • The framework separates thematic variation from genre-form control by combining diverse human-authored story prompts with manually curated genre attributes.It targets systematic genre coverage, genre-form fidelity, and specification diversity.Source: §1 · p. 2

Contribution

  • 50K instances span 13 creative-writing genres in the resulting Multi-Genre Collection.The collection is designed to scale creative-writing data beyond story-centric sources.Source: §1 · p. 2

Results

  • Models fine-tuned on the Multi-Genre Collection consistently outperform base models, LongWriter-glm4-9B, and models trained on existing writing corpora across popular benchmarks.Increasing genre coverage also improves output novelty relative to story-only training.Source: §1 · p. 2

2 Attribute-Guided Genre Expansion

The Multi-Genre Collection is constructed through attribute-guided genre expansion, combining human-authored story seeds with genre-specific attributes to generate genre-faithful creative writing pairs. A three-stage pipeline produces a 50K-instance English corpus spanning 13 creative genres, with quality filtering and balanced genre coverage.

Framework

  • The framework expands beyond story-centric datasets by generating queries for each target genre through attribute-guided genre expansion.It starts from a human-authored story-centric seed pool rather than directly collecting prompts for every genre.Source: §2 · pp. 2–3
  • The pipeline separates thematic breadth from genre-form control: seed examples provide topical and stylistic variation, while attributes enforce structural conventions.Genre attributes include dimensions such as rap rhyme scheme and flow, or TV-series narrative arc and character development.Source: §2 · p. 3

Pipeline

  • Three stages sample n=5 thematic seed pairs, sample genre attributes, and synthesize attribute-guided query-response pairs with LLMs.The synthesis stage generates five queries per template instantiation, uses verbalized sampling for diversity, and filters low-quality responses with an independent LLM judge.Source: §2 · p. 3

Outcomes

  • The resulting Multi-Genre Collection contains 50K instances with balanced coverage across 13 creative genres.The “Etc.” category includes long-tail formats such as diary entries, comic scripts, and interactive fiction; t-SNE visualization is used to assess semantic distinctiveness.Source: §2 · p. 3

3 Experiments

Experiments show that SFT on the Multi-Genre Collection improves base models across out-of-distribution and in-distribution creative-writing evaluations. Gains remain robust against independent judges and human evaluation, while genre diversity improves novelty and format-specific writing quality.

Models and Benchmarks

  • Three base models are fine-tuned with LoRA-based SFT and evaluated on Arena Hard, WritingBench, and a held-out Multi-Genre test set.The models are Llama-3.1-8B-Instruct, EXAONE-3.5-7.8B-Instruct, and Qwen3-8B.Source: §3.1 · p. 4

Overall Performance

  • All three base models improve substantially after SFT, with consistent out-of-distribution gains and balanced coverage across 13 genres.The largest out-of-distribution gains occur for Qwen3-8B.Source: §3.2 · p. 4
  • All fine-tuned models outperform the writing-specialized LongWriter-glm4-9B baseline, showing the benefit of diverse genre-specific data over narrow specialization.This comparison is reported as a consistent result across the evaluated models.Source: §3.2 · p. 4

Comparison with Existing Writing Datasets

  • Qwen3-8B trained on the Multi-Genre Collection substantially outperforms models trained on DeepWriting-20k and LongWriter-6k across all benchmarks.The comparison uses randomly sampled 2K examples from each dataset for SFT, with particularly large gains on Arena Hard.Source: §3.2 · p. 4

Impact of Genre Diversity

  • NoveltyBench Distinct scores rise monotonically as training genres increase from 0 to 13, indicating that genre diversity enhances output distinctness.The metric estimates distinctness by clustering semantically and functionally equivalent generations.Source: §3.2 · p. 4, Table 3 · p. 4

Robustness and Qualitative Analysis

  • Independent-judge evaluation and human evaluation confirm that fine-tuned variants outperform their base models, while qualitative outputs show improved format compliance, specificity, and coherence.The qualitative example adds titles, named entities, a recurring character, specific spells, and section-level timestamps.Source: §3.2 · pp. 4–5, Figure 6 · p. 5

4 Related Work

Prior creative-writing datasets have largely centered on story generation, with established resources covering prompt–story pairs, short commonsense narratives, and preference-based evaluation. Other work has addressed individual genres separately rather than broad genre coverage.

4 Related Work

  • Early creative-writing datasets focused almost exclusively on story generation.This included resources designed for narrative understanding and story-quality evaluation.Source: §4 · p. 5
  • WritingPrompts provided diverse prompt–story pairs sourced from Reddit.It was introduced as a dataset for creative writing research.Source: §4 · p. 5
  • ROCStories offered five-sentence commonsense stories for narrative understanding.The dataset targeted concise narrative examples rather than multiple creative genres.Source: §4 · p. 5
  • LitBench introduced preference labels for evaluating story quality and writing personalization.Its evaluation focus remained centered on stories and writer-specific preferences.Source: §4 · p. 5
  • Prior work beyond stories targeted individual genres in isolation.The passage begins an example with SongCompos, but does not provide further details.Source: §4 · p. 5

5 Conclusion

The paper introduces the 50K-instance Multi-Genre Collection spanning 13 genres, built through attribute-guided genre expansion. Experiments show that models trained on it outperform base models, writing-specialized baselines, and existing writing datasets.

Dataset

  • The Multi-Genre Collection contains 50K instances across 13 genres.It is the paper’s introduced dataset.Source: §5 · p. 5

Method

  • Attribute-guided genre expansion transfers thematic diversity from human-authored prompts while enforcing genre-specific constraints with manually curated attributes.The framework combines diverse thematic seeds with explicit genre control.Source: §5 · p. 5

Findings

  • Models fine-tuned on the collection substantially outperform both base models and writing-specialized baselines.This is one of the paper’s three reported experimental findings.Source: §5 · p. 5
  • The collection consistently outperforms existing writing datasets.This is another of the paper’s three reported experimental findings.Source: §5 · p. 5

GenAI Usage Disclosure

The manuscript was written by the authors, with ChatGPT used only to refine style, clarity, and grammar. Large language models also serve as components of the data-construction pipeline for query generation, response generation, filtering, and quality scoring.

  • ChatGPT was used solely to refine the manuscript’s style, clarity, and grammar, not for ideation, claim generation, or experimental design.Source: GenAI Usage Disclosure · p. 5
  • The data-construction pipeline uses GPT-5-mini for filtering and query generation, Qwen3-235B-A22B-Thinking for response generation, and Qwen3-30B-A3B-Instruct as a quality-scoring LLM-as-a-judge.Source: GenAI Usage Disclosure · p. 5

Related papers

Loading 2608.13947v1…