Elamur Research⌕
☰
home / papers / 2026 / ChatAnime: Towards User-Centered Emotional Support in LLM-based Virtual Character Chat

ChatAnime: Towards User-Centered Emotional Support in LLM-based Virtual Character Chat

Paper · 2026 · Lanlan Qiu, Sophia Xiao Pu, Yeqi Feng et al.

Summary

Users of virtual character platforms such as Character.AI increasingly seek emotional comfort from role-playing agents. Prior role-playing research, however, mostly checks whether a character stays consistent in fictional or game settings, not whether it helps real users with real problems. The authors define Emotionally Supportive Role-Playing (ESRP) and build ChatAnime, a Chinese dataset in which 20 popular anime characters answer everyday emotional struggles.

GPT-4o generated scenarios by combining user personas, emotional states and locations, and the authors manually picked 60 of them. Two-turn replies came from 40 selected anime fans and 10 LLMs, and human raters then scored them on 9 metrics plus diversity. The metrics fall into three groups: basic dialogue quality, role-playing and emotional support.

The strongest models, led by DeepSeek-V3, outscored fans on role-playing and support. Humans, however, gave far more varied answers. The authors also release a fine-tuned judge model that agrees with human ratings better than commercial LLMs do. The resource matters for evaluating companion-style chatbots that focus on user well-being.

Key points

  • ChatAnime scale: 20 characters, 60 scenario questions, 2,400 human-written and 24,000 LLM-generated responses, and over 132,000 human annotations.
  • Scenario design: 288 candidate questions came from 4 personas (student, employee, freelancer, self-employed), 9 emotions and 4 locations per persona. The 40 fan annotators were chosen from roughly 300 applicants.
  • Human vs. model scores: DeepSeek-V3 had the best average human score (4.10), compared with 3.95 for human fans.
  • Diversity: Humans led on diversity (3.90). GPT-4.1 was the most diverse model (3.57), and DeepSeek-V3 the least (2.93), which hints at a trade-off between quality and variety. Agreement between raters, measured by weighted Kendall's tau, was 0.55.
  • Judge model and limitations: The judge is Qwen3-8B fine-tuned with LoRA (r=8). It reaches 73% accuracy and above 96% when a one-point deviation is allowed, versus 43–54% for Gemini-2.5-Flash-Preview, GPT-4.1 and Qwen-Max. The data are limited to Chinese speakers and two-turn dialogues.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Tsinghua UniversityBeijing · China333 works · 87,019 citations

 Fields, topics and concepts

📚 Shelves 7The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksPsychology · 1,552 worksExperimental and Cognitive Psychology · 870 worksSocial Psychology · 306 works

🧭 Topics 1Research subjects and languages the work is about.Sentiment & emotion · 972 works

Cite

Cite as Lanlan Qiu, Sophia Xiao Pu, Yeqi Feng, Wenchang Gao and Tianxing He. 2026. ChatAnime: Towards User-Centered Emotional Support in LLM-based Virtual Character Chat. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.2179

@inproceedings{qiu2026chatanime,
  title = {{ChatAnime: Towards User-Centered Emotional Support in LLM-based Virtual Character Chat}},
  author = {Lanlan Qiu, Sophia Xiao Pu, Yeqi Feng, Wenchang Gao and Tianxing He},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.2179},
  url = {https://doi.org/10.18653/v1/2026.acl-long.2179}
}

References

35 references, 14 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Nuo Chen, Yan Wang, Haiyun Jiang, Deng Cai, Yuhan Li, Ziyang Chen, Longyue Wang, and Jia Li. 2023. Large language models meet harry potter: A dataset for aligning dialogue agents with characters. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 8506–8520, Singapore. As…
  2. Team GLM, Aohan Zeng, Bin Xu, Bowen Wang, Chenhui Zhang, Da Yin, Dan Zhang, Diego Rojas, Guanyu Feng, Hanlin Zhao, and 1 others. 2024. Chatglm: A family of large language models from glm-130b to glm-4 all tools. arXiv preprint arXiv:2406.12793.
  3. Google. 2025. Gemini 2.5 flash. find
  4. Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, and 1 others. 2022. Lora: Low-rank adaptation of large language models. ICLR, 1(2):3.
  5. Yining Hua, Fenglin Liu, Kailai Yang, Zehan Li, Hongbin Na, Yi han Sheu, Peilin Zhou, Lauren V. Moran, Sophia Ananiadou, David A. Clifton, Andrew Beam, and John Torous. 2025. Large language models in mental health care: a scoping review. Preprint, arXiv:2401.02984. iResearch. 2021. Anime reports. — [arXiv:2401.02984](https://arxiv.org/abs/2401.02984) find
  6. Haoan Jin, Siyuan Chen, Dilawaier Dilixiati, Yewei Jiang, Mengyue Wu, and Kenny Q Zhu. 2023. Psyeval: A suite of mental health related tasks for evaluating large language models. arXiv preprint arXiv:2311.09189. — [arXiv:2311.09189](https://arxiv.org/abs/2311.09189) find
  7. Cheng Li, Ziang Leng, Chenxi Yan, Junyi Shen, Hao Wang, Weishi Mi, Yaying Fei, Xiaoyang Feng, Song Yan, HaoSheng Wang, and 1 others. 2023. Chatharuhi: Reviving anime character in reality via large language model. arXiv preprint arXiv:2308.09597.
  8. Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, and 1 others. 2024. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437. find
  9. Alibaba. 2024. Xingchen-plus-v2. find
  10. Anthropic. 2025. Claude sonnet 4. find
  11. Lennart Brocki, George C. Dyer, Anna Gładka, and Neo Christopher Chung. 2023. Deep learning mental health dialogue system. Preprint, arXiv:2301.09412. — [arXiv:2301.09412](https://arxiv.org/abs/2301.09412) find
  12. ByteDance. 2025a. Doubao-1.5-pro. find
  13. ByteDance. 2025b. Doubao-rp. find
  14. Jiangjie Chen, Xintao Wang, Rui Xu, Siyu Yuan, Yikai Zhang, Wei Shi, Jian Xie, Shuang Li, Ruihan Yang, Tinghui Zhu, and 1 others. 2024a. From persona to personalization: A survey on role-playing language agents. arXiv preprint arXiv:2404.18231.
  15. Nuo Chen, Yan Wang, Yang Deng, and Jia Li. 2024b. The oscars of ai theater: A survey on role-playing with language models. arXiv preprint arXiv:2407.11484. — [arXiv:2407.11484](https://arxiv.org/abs/2407.11484) find
  16. June M. Liu, Donghao Li, He Cao, Tianhe Ren, Zeyi Liao, and Jiamin Wu. 2023. Chatcounselor: A large language models for mental health support. Preprint, arXiv:2309.15461. — [arXiv:2309.15461](https://arxiv.org/abs/2309.15461) find
  17. Siyang Liu, Chujie Zheng, Orianna Demasi, Sahand Sabour, Yu Li, Zhou Yu, Yong Jiang, and Minlie Huang. 2021. Towards emotional support dialog systems. Preprint, arXiv:2106.01144.
  18. Keming Lu, Bowen Yu, Chang Zhou, and Jingren Zhou. 2024. Large language models are superpositions of all characters: Attaining arbitrary role-play via self-alignment. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 7828–7840,…
  19. MiniMax. 2024. Minimax-abab6.5s. and Le Sun. 2025. Rmtbench: Benchmarking llms through multi-turn user-centric role-playing. Preprint, arXiv:2507.20352. find
  20. OpenAI. 2025. Model - openai api. find
  21. Yunfan Shao, Linyang Li, Junqi Dai, and Xipeng Qiu. 2023. Character-llm: A trainable agent for role-playing. Preprint, arXiv:2310.10158.
  22. Yuchong Sun, Che Liu, Kun Zhou, Jinwen Huang, Ruihua Song, Wayne Xin Zhao, Fuzheng Zhang, Di Zhang, and Kun Gai. 2024. Parrot: Enhancing multi-turn instruction following for large language models. Preprint, arXiv:2310.07301. find
  23. Yu-Min Tseng, Yu-Chao Huang, Teng-Yun Hsiao, Wei-Lin Chen, Chao-Wei Huang, Yu Meng, and Yun-Nung Chen. 2024. Two tales of persona in llms: A survey of role-playing and personalization. arXiv preprint arXiv:2406.01171.
  24. Quan Tu, Shilong Fan, Zihang Tian, Tianhao Shen, Shuo Shang, Xin Gao, and Rui Yan. 2024. CharacterEval: A Chinese benchmark for role-playing conversational agent evaluation. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1183…
  25. Ming Wang, Peidong Wang, Lin Wu, Xiaocui Yang, Daling Wang, Shi Feng, Yuxin Chen, Bixuan Wang, and Yifei Zhang. 2025a. Annaagent: Dynamic evolution agent system with multi-session memory for realistic seeker simulation. arXiv preprint arXiv:2506.00551. find
  26. Xintao Wang, Heng Wang, Yifei Zhang, Xinfeng Yuan, Rui Xu, Jen tse Huang, Siyu Yuan, Haoran Guo, Jiangjie Chen, Shuchang Zhou, Wei Wang, and Yanghua Xiao. 2025b. Coser: Coordinating llm-based persona simulation of established roles. Preprint, arXiv:2502.09082. — [arXiv:2502.09082](https://arxiv.org/abs/2502.09082) find
  27. Xintao Wang, Yunze Xiao, Jen tse Huang, Siyu Yuan, Rui Xu, Haoran Guo, Quan Tu, Yaying Fei, Ziang Leng, Wei Wang, Jiangjie Chen, Cheng Li, and Yanghua Xiao. 2024a. Incharacter: Evaluating personality fidelity in role-playing agents through psychological interviews. Preprint, arXiv:2310.17976.
  28. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, and 1 others. 2025. Qwen3 technical report. arXiv preprint arXiv:2505.09388.
  29. Dingbo Yuan, Yipeng Chen, Guodong Liu, Chenchen Li, Chengfu Tang, Dongxu Zhang, Zhenkui Wang, Xudong Wang, and Song Liu. 2025. Dmt-rolebench: A dynamic multi-turn dialogue based benchmark for role-playing evaluation of large language model and agent. In Proceedings of the AAAI Conference on Artifici… find
  30. Chenhao Zhang, Renhao Li, Minghuan Tan, Min Yang, Jingwei Zhu, Di Yang, Jiahao Zhao, Guancheng Ye, Chengming Li, and Xiping Hu. 2024. Cpsycoun: A report-based multi-turn dialogue reconstruction and evaluation framework for chinese psychological counseling. Preprint, arXiv:2405.16433. find
  31. Haiquan Zhao, Lingyu Li, Shisong Chen, Shuqi Kong, Jiaan Wang, Kexin Huang, Tianle Gu, Yixu Wang, Jian Wang, Liang Dandan, Zhixu Li, Yan Teng, Yanghua Xiao, and Yingchun Wang. 2024. ESC-eval: Evaluating emotion support conversations in large language models. In Proceedings of the 2024 Conference on… find
  32. Jinfeng Zhou, Zhuang Chen, Dazhen Wan, Bosi Wen, Yi Song, Jifan Yu, Yongkang Huang, Libiao Peng, Jiaming Yang, Xiyao Xiao, Sahand Sabour, Xiaohan Zhang, Wenjing Hou, Yijia Zhang, Yuxiao Dong, Jie Tang, and Minlie Huang. 2023. Characterglm: Customizing chinese conversational ai characters with large…
  33. Zekun Moore Wang, Zhongyuan Peng, Haoran Que, Jiaheng Liu, Wangchunshu Zhou, Yuhan Wu, Hongcheng Guo, Ruitong Gan, Zehao Ni, Jian Yang, Man Zhang, Zhaoxiang Zhang, Wanli Ouyang, Ke Xu, Stephen W. Huang, Jie Fu, and Junran Peng. 2024b. Rolellm: Benchmarking, eliciting, and enhancing role-playing abil…
  34. Zongsheng Wang, Kaili Sun, Bowen Wu, Qun Yu, Ying Li, and Baoxun Wang. 2025c. Raiden-r1: Improving role-awareness of llms via grpo with verifiable reward. Preprint, arXiv:2505.10218. — [arXiv:2505.10218](https://arxiv.org/abs/2505.10218) find
  35. Hao Xiang, Tianyi Tang, Yang Su, Bowen Yu, An Yang, Fei Huang, Yichang Zhang, Yaojie Lu, Hongyu Lin, Xianpei Han, Jingren Zhou, Junyang Lin, find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/qiu-2026-chatanime-user-centered-emotional-support/