Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
October 20, 2025Open Access

AudioRole: An Audio Dataset for Character Role-Playing in Large Language Models

View Full Paper
Ask AI
Bookmark
Share

Authors

WLWenyu LiXJXiaoqi JiaoYCYi Chang

Discussion

Loading...

Member takes

Overview

This dataset features synchronized audio-text pairs and character-grounded dialogues, suggesting its potential for advanced training in language modeling.

Key Points

  • AudioRole provides over 1M character-grounded dialogues from 13 TV series, enabling enhanced role-playing in models.
  • The ARP-Model trained on AudioRole achieves an Acoustic Personalization score of 0.31, significantly surpassing previous models.
  • Using a dual-aspect evaluation framework, AudioRole uniquely assesses response quality alongside role fidelity.
  • This dataset equips researchers with a vital resource, crucial for pushing the boundaries of audio-grounded role-playing.

Cite This Study

Li et al. (2025) studied this question.

synapsesocial.com/papers/68f6379bb481a140a36cf418https://doi.org/10.48550/arxiv.2509.23435
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Video2Roleplay: A Multimodal Dataset and Framework for Video-Guided Role-playing Agents2025
  2. 2OmniCharacter++: Towards Comprehensive Benchmark for Realistic Role-Playing Agents2026
  3. 3Capturing Minds, Not Just Words: Enhancing Role-Playing Language Models with Personality-Indicative Data2024
  4. 4OmniCharacter: Towards Immersive Role-Playing Agents with Seamless Speech-Language Personality Interaction2025
  5. 5Prompt Framework for Role-playing: Generation and Evaluation2024 · 2 citations