pith. sign in

arxiv: 2606.25632 · v1 · pith:TBQUEE5Jnew · submitted 2026-06-24 · 💻 cs.CL · cs.AI

Staying In Character: Perspective-Bounded Memory For Book-Based Role-Playing Agents

Xushuo Tang , Junhe Zhang , Zihan Yang , Yifu Tang , Sichao Li , Longbin Lai , Zhengyi Yang This is my paper

Pith reviewed 2026-06-25 20:54 UTC · model grok-4.3

classification 💻 cs.CL cs.AI
keywords role-playing agentsmemory architectureknowledge boundariesLLM agentscharacter consistencybook-based narrativesperspective bounding
0
0 comments X

The pith

A three-layer memory architecture keeps book characters from using facts outside their own perspective.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents REVERIEMEM as a memory system for LLM agents that role-play characters drawn from novels. It separates first-person scene recollections, facts marked by who can see them, and patterns of speech and action that shift with situation. The design targets two problems: agents reciting information their character could not know, and agents falling into repetitive voice from static profiles. New benchmarks show the system raises the rate at which agents stay inside character knowledge limits and produces narratives preferred by judges on multiple story dimensions.

Core claim

REVERIEMEM stores episodic memories as first-person scene records, semantic facts tagged by visibility to the character, and personality patterns that vary by context, allowing agents to respect the knowledge boundaries of book characters during long narrative interactions.

What carries the argument

Three-layer memory architecture that stores first-person scenes, visibility-tagged facts, and situation-dependent personality patterns to enforce perspective bounds.

Load-bearing premise

The new KBF-QA questions and BOOKWORLD protocol measure only the effect of the memory layers rather than differences in the underlying model or retrieval code.

What would settle it

An experiment that shows an agent using the three-layer memory still answers questions with facts visible only to other characters in the same novel.

Figures

Figures reproduced from arXiv: 2606.25632 by Junhe Zhang, Longbin Lai, Sichao Li, Xushuo Tang, Yifu Tang, Zhengyi Yang, Zihan Yang.

Figure 1
Figure 1. Figure 1: Two OOC failures in long-narrative role playing: [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overview of REVERIEMEM. Given a query on character c, the system runs three phases described in §3: 1) Source-to-Memory extracts per-scene text, dialogue, and emotion data via an LLM and constructs a perspective-bounded three-layer memory for each focus character; 2) CLS Memory Collaborative Reasoning anchors on the Episodic Layer scene memory Sc, which both grounds and informs SELF-PROBE to iteratively ex… view at source ↗
Figure 3
Figure 3. Figure 3: Two example KBF-QA items from Dracula, posed to Dr. John Seward. 15 [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
read the original abstract

Recent LLM role-playing systems build character agents from novels by extracting characters, scenes, and relations. Yet long-narrative role-playing suffers from two failures: Factual Overreach, where shared retrieval or parametric memory lets a character use facts outside its perspective, and Stylistic Monotony, where profile descriptions flatten a character into a fixed voice. To address these failures, we propose REVERIEMEM, a three-layer memory architecture for book-based character agents. The episodic layer stores first-person scene memories; the semantic layer stores visibility-tagged facts; and the personality layer stores situation-dependent speech and behaviour patterns. For evaluation, we construct KBF-QA, a 4,386-question benchmark over eight novels for testing knowledge boundaries. REVERIEMEM improves Knowledge Boundary Fidelity by 34.6 percentage points over the strongest prior method. On BOOKWORLD's five-dimension pairwise narrative protocol, REVERIEMEM achieves a ~ 79% win rate, suggesting that perspective-bounded memory improves both boundary fidelity and character-grounded narrative generation.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper proposes REVERIEMEM, a three-layer perspective-bounded memory architecture (episodic first-person scenes, semantic visibility-tagged facts, personality situation-dependent patterns) for LLM role-playing agents extracted from novels. It introduces the KBF-QA benchmark (4,386 questions over eight novels) to measure knowledge boundary fidelity and evaluates narrative quality via BOOKWORLD's five-dimension pairwise protocol, claiming a 34.6 percentage point KBF gain over the strongest baseline and a ~79% win rate.

Significance. If the quantitative claims are supported by controlled experiments that isolate the memory architecture, the work would offer a concrete mechanism for reducing factual overreach and stylistic flattening in book-derived agents, together with a reusable benchmark for perspective adherence. The layered design and benchmark construction represent potentially useful contributions to character consistency research.

major comments (2)
  1. [Abstract] Abstract: The headline claims of a 34.6pp KBF improvement and ~79% win rate are load-bearing for the central thesis yet supply no experimental protocol, baseline descriptions, LLM/prompt/retrieval controls, or statistical tests. Without these, it is impossible to determine whether the gains are attributable to the three-layer design rather than confounds.
  2. [Evaluation] Evaluation section (inferred from abstract claims): The KBF-QA and BOOKWORLD protocols are presented as isolating the effect of perspective-bounded memory, but no evidence is given that baselines share identical base models, system prompts, and retrieval implementations. This control is required to support the causal attribution stated in the abstract.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. The comments highlight the need for greater transparency in experimental controls to support causal claims about the memory architecture. We address each point below and will revise the manuscript accordingly to strengthen the presentation of results.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The headline claims of a 34.6pp KBF improvement and ~79% win rate are load-bearing for the central thesis yet supply no experimental protocol, baseline descriptions, LLM/prompt/retrieval controls, or statistical tests. Without these, it is impossible to determine whether the gains are attributable to the three-layer design rather than confounds.

    Authors: We agree that the abstract, due to length constraints, omits key experimental details. The full protocols, baseline descriptions (including RAG, profile-only, and memory-augmented variants), LLM choices (GPT-4o and Llama-3-70B), prompt templates, retrieval implementations, and statistical significance tests (paired t-tests with p<0.01) are documented in Section 4 (Evaluation) and Appendix B. In the revision we will add one sentence to the abstract summarizing the shared experimental controls to make the claims more self-contained while preserving conciseness. revision: yes

  2. Referee: [Evaluation] Evaluation section (inferred from abstract claims): The KBF-QA and BOOKWORLD protocols are presented as isolating the effect of perspective-bounded memory, but no evidence is given that baselines share identical base models, system prompts, and retrieval implementations. This control is required to support the causal attribution stated in the abstract.

    Authors: All methods were run with identical base models, system prompts, and retrieval pipelines; only the memory component differed. This is stated in Section 4.1 (“Experimental Setup”) and Table 2. To address the concern directly, the revision will include an explicit paragraph confirming these controls and noting that any differences in performance are therefore attributable to the three-layer perspective-bounded design rather than implementation variance. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical architecture evaluated on external benchmarks

full rationale

The paper proposes REVERIEMEM as a three-layer memory design and reports empirical gains on KBF-QA (4,386 questions) and BOOKWORLD protocol. No equations, fitted parameters, self-definitional loops, or load-bearing self-citations appear in the provided text. Claims rest on benchmark comparisons rather than any derivation that reduces to its own inputs by construction. The evaluation protocol is presented as an independent test of perspective-bounded memory, with no indication that results are forced by the architecture definition itself.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no information on free parameters, axioms, or invented entities; full text unavailable for audit.

pith-pipeline@v0.9.1-grok · 5731 in / 1123 out tokens · 33335 ms · 2026-06-25T20:54:13.792586+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

47 extracted references · 13 canonical work pages

  1. [1]

    2025 , eprint=

    BookWorld: From Novels to Interactive Agent Societies for Creative Story Generation , author=. 2025 , eprint=

    2025

  2. [2]

    HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models , url =

    Guti\'. HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models , url =. Advances in Neural Information Processing Systems , doi =

  3. [4]

    , author=

    Why there are complementary learning systems in the hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory. , author=. Psychological review , volume=. 1995 , publisher=

    1995

  4. [5]

    , author=

    The construction of autobiographical memories in the self-memory system. , author=. Psychological review , volume=. 2000 , publisher=

    2000

  5. [6]

    Narrative Identity , volume =

    Mcadams, Dan and McLean, Kate , year =. Narrative Identity , volume =. Current Directions in Psychological Science , doi =

  6. [7]

    Character- LLM : A Trainable Agent for Role-Playing

    Shao, Yunfan and Li, Linyang and Dai, Junqi and Qiu, Xipeng. Character- LLM : A Trainable Agent for Role-Playing. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023

    2023

  7. [9]

    2025 , eprint=

    OpenCharacter: Training Customizable Role-Playing LLMs with Large-Scale Synthetic Personas , author=. 2025 , eprint=

    2025

  8. [10]

    2025 , editor =

    Wang, Xintao and Wang, Heng and Zhang, Yifei and Yuan, Xinfeng and Xu, Rui and Huang, Jen-Tse and Yuan, Siyu and Guo, Haoran and Chen, Jiangjie and Zhou, Shuchang and Wang, Wei and Xiao, Yanghua , booktitle =. 2025 , editor =

    2025

  9. [15]

    AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors , url =

    Chen, Weize and Su, Yusheng and Zuo, Jingwei and Yang, Cheng and Yuan, Chenfei and Chan, Chi-Min and Yu, Heyang and Lu, Yaxi and Hung, Yi-Hsin and Qian, Chen and Qin, Yujia and Cong, Xin and Xie, Ruobing and Liu, Zhiyuan and Sun, Maosong and Zhou, Jie , booktitle =. AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors , url =

  10. [18]

    2026 , url=

    RoleArena: A Multi-Agent Role-Playing Environment for Long Multi-Turn Dialogues with Autonomous Plot Progression , author=. 2026 , url=

    2026

  11. [19]

    Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks , url =

    Lewis, Patrick and Perez, Ethan and Piktus, Aleksandra and Petroni, Fabio and Karpukhin, Vladimir and Goyal, Naman and K\". Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks , url =. Advances in Neural Information Processing Systems , editor =

  12. [20]

    International conference on learning representations , volume=

    Self-rag: Learning to retrieve, generate, and critique through self-reflection , author=. International conference on learning representations , volume=

  13. [21]

    RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval , url =

    Sarthi, Parth and Abdullah, Salman and Tuli, Aditi and Khanna, Shubh and Goldie, Anna and Manning, Christopher , booktitle =. RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval , url =

  14. [24]

    2024 , eprint=

    ERABAL: Enhancing Role-Playing Agents through Boundary-Aware Learning , author=. 2024 , eprint=

    2024

  15. [26]

    NarrativePlay: Interactive Narrative Understanding

    Runcong Zhao and Wenjia Zhang and Jiazheng Li and Lixing Zhu and Yanran Li and Yulan He and Lin Gui. NarrativePlay: Interactive Narrative Understanding. EACL 2024 - 18th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings of System Demonstrations. 2024

    2024

  16. [27]

    2026 , eprint=

    OpenAI GPT-5 System Card , author=. 2026 , eprint=

    2026

  17. [28]

    Qwen3.5: Accelerating Productivity with Native Multimodal Agents , url =

    Qwen Team , month =. Qwen3.5: Accelerating Productivity with Native Multimodal Agents , url =

  18. [29]

    2025 , eprint=

    DeepSeek-V3 Technical Report , author=. 2025 , eprint=

    2025

  19. [30]

    Akari Asai, Zeqiu Wu, Yizhong Wang, Avi Sil, and Hannaneh Hajishirzi. 2024. Self-rag: Learning to retrieve, generate, and critique through self-reflection. In International conference on learning representations, volume 2024, pages 9112--9141

    2024

  20. [31]

    Jing Chen, Xinyu Zhu, Cheng Yang, Chufan Shi, Yadong Xi, Yuxiang Zhang, Junjie Wang, Jiashu Pu, Tian Feng, Yujiu Yang, and Rongsheng Zhang. 2024 a . https://doi.org/10.18653/v1/2024.findings-emnlp.474 H o LLM wood: Unleashing the creativity of large language models in screenwriting via role playing . In Findings of the Association for Computational Lingui...

    work page doi:10.18653/v1/2024.findings-emnlp.474 2024

  21. [32]

    Weize Chen, Yusheng Su, Jingwei Zuo, Cheng Yang, Chenfei Yuan, Chi-Min Chan, Heyang Yu, Yaxi Lu, Yi-Hsin Hung, Chen Qian, Yujia Qin, Xin Cong, Ruobing Xie, Zhiyuan Liu, Maosong Sun, and Jie Zhou. 2024 b . https://proceedings.iclr.cc/paper_files/paper/2024/file/578e65cdee35d00c708d4c64bce32971-Paper-Conference.pdf Agentverse: Facilitating multi-agent colla...

    2024

  22. [33]

    Martin A Conway and Christopher W Pleydell-Pearce. 2000. The construction of autobiographical memories in the self-memory system. Psychological review, 107(2):261

    2000

  23. [34]

    DeepSeek-AI, Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, Damai Dai, Daya Guo, Dejian Yang, Deli Chen, Dongjie Ji, Erhang Li, Fangyun Lin, Fucong Dai, and 181 others. 2025. https://arxiv.org/abs/2412.19437 Deepseek-v3 technical report . Preprint, arXiv:2412.19437

    Pith/arXiv arXiv 2025

  24. [35]

    Qiming Feng, Qiujie Xie, Xiaolong Wang, Qingqiu Li, Yuejie Zhang, Rui Feng, Tao Zhang, and Shang Gao. 2025. https://doi.org/10.18653/v1/2025.naacl-long.316 E mo C haracter: Evaluating the emotional fidelity of role-playing agents in dialogues . In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computationa...

    work page doi:10.18653/v1/2025.naacl-long.316 2025

  25. [36]

    Bernal Jim\' e nez Guti\' e rrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga, and Yu Su. 2024. https://doi.org/10.52202/079017-1902 Hipporag: Neurobiologically inspired long-term memory for large language models . In Advances in Neural Information Processing Systems, volume 37, pages 59532--59569. Curran Associates, Inc

    work page doi:10.52202/079017-1902 2024

  26. [37]

    u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt\

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K\" u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt\" a schel, Sebastian Riedel, and Douwe Kiela. 2020. https://proceedings.neurips.cc/paper_files/paper/2020/file/6b493230205f780e1bc26945df7481e5-Paper.pdf Retrieval-augmented generation for knowledge-intens...

    2020

  27. [38]

    Moxin Li, Yong Zhao, Wenxuan Zhang, Shuaiyi Li, Wenya Xie, See-Kiong Ng, Tat-Seng Chua, and Yang Deng. 2025. https://doi.org/10.18653/v1/2025.acl-long.256 Knowledge boundary of large language models: A survey . In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5131--5157, Vienna, Aust...

    work page doi:10.18653/v1/2025.acl-long.256 2025

  28. [39]

    Wenhao Liu, Siyu An, Junru Lu, Muling Wu, Tianlong Li, Xiaohua Wang, Changze Lv, Xiaoqing Zheng, Di Yin, Xing Sun, and Xuanjing Huang. 2025. https://doi.org/10.18653/v1/2025.findings-acl.311 Tell me what you don ' t know: Enhancing refusal capabilities of role-playing agents via representation space analysis and editing . In Findings of the Association fo...

    work page doi:10.18653/v1/2025.findings-acl.311 2025

  29. [40]

    Dan Mcadams and Kate McLean. 2013. https://doi.org/10.1177/0963721413475622 Narrative identity . Current Directions in Psychological Science, 22:233--238

    work page doi:10.1177/0963721413475622 2013

  30. [41]

    James L McClelland, Bruce L McNaughton, and Randall C O'Reilly. 1995. Why there are complementary learning systems in the hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory. Psychological review, 102(3):419

    1995

  31. [42]

    Generative agents: Interactive simulacra of human behavior,

    Joon Sung Park, Joseph O'Brien, Carrie Jun Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. 2023. https://doi.org/10.1145/3586183.3606763 Generative agents: Interactive simulacra of human behavior . In Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology, UIST '23, New York, NY, USA. Association for Com...

    work page doi:10.1145/3586183.3606763 2023

  32. [43]

    Yiting Ran, Xintao Wang, Tian Qiu, Jiaqing Liang, Yanghua Xiao, and Deqing Yang. 2025. https://arxiv.org/abs/2504.14538 Bookworld: From novels to interactive agent societies for creative story generation . Preprint, arXiv:2504.14538

    arXiv 2025

  33. [44]

    Parth Sarthi, Salman Abdullah, Aditi Tuli, Shubh Khanna, Anna Goldie, and Christopher Manning. 2024. https://proceedings.iclr.cc/paper_files/paper/2024/file/8a2acd174940dbca361a6398a4f9df91-Paper-Conference.pdf Raptor: Recursive abstractive processing for tree-organized retrieval . In International Conference on Learning Representations, volume 2024, page...

    2024

  34. [45]

    Yunfan Shao, Linyang Li, Junqi Dai, and Xipeng Qiu. 2023. https://aclanthology.org/2023.emnlp-main.814/ Character- LLM : A trainable agent for role-playing . In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 13153--13187, Singapore. Association for Computational Linguistics

    2023

  35. [46]

    Aaditya Singh, Adam Fry, Adam Perelman, Adam Tart, Adi Ganesh, Ahmed El-Kishky, Aidan McLaughlin, Aiden Low, AJ Ostrow, Akhila Ananthram, Akshay Nathan, Alan Luo, Alec Helyar, Aleksander Madry, Aleksandr Efremov, Aleksandra Spyra, Alex Baker-Whitcomb, Alex Beutel, Alex Karpenko, and 467 others. 2026. https://arxiv.org/abs/2601.03267 Openai gpt-5 system ca...

    Pith/arXiv arXiv 2026

  36. [47]

    Yihong Tang, Jiao Ou, Che Liu, Fuzheng Zhang, Di Zhang, and Kun Gai. 2024. https://arxiv.org/abs/2409.14710 Erabal: Enhancing role-playing agents through boundary-aware learning . Preprint, arXiv:2409.14710

    arXiv 2024

  37. [48]

    Qwen Team. 2026. https://qwen.ai/blog?id=qwen3.5 Qwen3.5: Accelerating productivity with native multimodal agents

    2026

  38. [49]

    Yu-Min Tseng, Yu-Chao Huang, Teng-Yun Hsiao, Wei-Lin Chen, Chao-Wei Huang, Yu Meng, and Yun-Nung Chen. 2024. https://doi.org/10.18653/v1/2024.findings-emnlp.969 Two tales of persona in LLM s: A survey of role-playing and personalization . In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 16612--16631, Miami, Florida, USA. Ass...

    work page doi:10.18653/v1/2024.findings-emnlp.969 2024

  39. [50]

    Quan Tu, Shilong Fan, Zihang Tian, Tianhao Shen, Shuo Shang, Xin Gao, and Rui Yan. 2024. https://doi.org/10.18653/v1/2024.acl-long.638 C haracter E val: A C hinese 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 11836--118...

    work page doi:10.18653/v1/2024.acl-long.638 2024

  40. [51]

    Juyuan Wang, Rongchen Zhao, Wei Wei, Yufeng Wang, Mo Yu, Jie Zhou, Jin Xu, and Liyan Xu. 2026. https://doi.org/10.1609/aaai.v40i39.40644 Comorag: A cognitive-inspired memory-organized rag for stateful long narrative reasoning . Proceedings of the AAAI Conference on Artificial Intelligence, 40(39):33557–33565

    work page doi:10.1609/aaai.v40i39.40644 2026

  41. [52]

    Noah Wang, Z.y. 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, Wenhao Huang, Jie Fu, and Junran Peng. 2024 a . https://doi.org/10.18653/v1/2024.findings-acl.878 R ole LLM : Benchmarking, eliciting, and enhancing role-playing abilities of large lan...

    work page doi:10.18653/v1/2024.findings-acl.878 2024

  42. [53]

    Xiaoyang Wang, Hongming Zhang, Tao Ge, Wenhao Yu, Dian Yu, and Dong Yu. 2025 a . https://arxiv.org/abs/2501.15427 Opencharacter: Training customizable role-playing llms with large-scale synthetic personas . Preprint, arXiv:2501.15427

    arXiv 2025

  43. [54]

    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. 2025 b . https://proceedings.mlr.press/v267/wang25dk.html C o SER : Coordinating LLM -based persona simulation of established roles . In Proceedings of the 42nd International Conference on Machine Learn...

    2025

  44. [55]

    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. 2024 b . https://doi.org/10.18653/v1/2024.acl-long.102 I n C haracter: Evaluating personality fidelity in role-playing agents through psychological interviews . In Proceedings of the 62nd Annual Meeti...

    work page doi:10.18653/v1/2024.acl-long.102 2024

  45. [56]

    Haotian Xia, Hao Peng, Yunjia Qi, Bin Xu, Juanzi Li, Hou Lei, and Xiaozhi Wang. 2025. https://doi.org/10.1145/3746252.3761616 Storywriter: A multi-agent framework for long story generation . In Proceedings of the 34th ACM International Conference on Information and Knowledge Management, CIKM '25, page 6559–6563, New York, NY, USA. Association for Computin...

    work page doi:10.1145/3746252.3761616 2025

  46. [57]

    Qi Zhao, Jian Li, Yizhang Jin, Dongqi Liu, Hang Ding, Qi Song, and Xiangyang Li. 2026. https://openreview.net/forum?id=o1idr3SbjG Rolearena: A multi-agent role-playing environment for long multi-turn dialogues with autonomous plot progression

    2026

  47. [58]

    Runcong Zhao, Wenjia Zhang, Jiazheng Li, Lixing Zhu, Yanran Li, Yulan He, and Lin Gui. 2024. NarrativePlay: Interactive Narrative Understanding, pages 82--93. EACL 2024 - 18th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings of System Demonstrations. Association for Computational Linguistics (ACL). Publisher...

    2024