@inproceedings{zhang-etal-2026-dynsess,
title = "{D}yn{S}ess: Dynamic Session-Level Evaluation and Optimization Framework for Role-Playing Agents",
author = "Zhang, Rongsheng and
Tang, Jiji and
Ren, Junnan and
Bao, Zuyi and
Chen, Weijie and
Hu, Ruofan and
Lv, Tangjie and
Zhao, Zhou and
Zhang, Yan",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.689/",
pages = "14443--14461",
ISBN = "979-8-89176-462-0",
abstract = "Role-playing with large language models is fundamentally a *session-level* task, requiring agents to sustain character identity and interaction quality across extended multi-turn conversations. Yet existing evaluation and optimization methods remain largely turn-level, failing to capture long-horizon quality. We propose **DynSess**, a unified session-level framework for role-playing agents. **DynSess-Eval** scores complete dialogue sessions via rubrics targeting long-horizon behaviors. Leveraging its session-level rewards, we construct high-quality training trajectories through multi-turn lookahead search and train **DynSess-Character** with two complementary variants: DSPO (off-policy) and GSRPO (on-policy). Experiments show that DynSess-Eval aligns with human judgments substantially better than prior evaluators, and blind human evaluation further shows that DynSess-Character matches the strongest character model despite using substantially fewer parameters, while maintaining strong role consistency and interactive ability. The dataset and code are publicly available in https://github.com/Collab-Gen/DynSess."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-dynsess">
<titleInfo>
<title>DynSess: Dynamic Session-Level Evaluation and Optimization Framework for Role-Playing Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Rongsheng</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiji</namePart>
<namePart type="family">Tang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junnan</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zuyi</namePart>
<namePart type="family">Bao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weijie</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruofan</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tangjie</namePart>
<namePart type="family">Lv</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhou</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Role-playing with large language models is fundamentally a *session-level* task, requiring agents to sustain character identity and interaction quality across extended multi-turn conversations. Yet existing evaluation and optimization methods remain largely turn-level, failing to capture long-horizon quality. We propose **DynSess**, a unified session-level framework for role-playing agents. **DynSess-Eval** scores complete dialogue sessions via rubrics targeting long-horizon behaviors. Leveraging its session-level rewards, we construct high-quality training trajectories through multi-turn lookahead search and train **DynSess-Character** with two complementary variants: DSPO (off-policy) and GSRPO (on-policy). Experiments show that DynSess-Eval aligns with human judgments substantially better than prior evaluators, and blind human evaluation further shows that DynSess-Character matches the strongest character model despite using substantially fewer parameters, while maintaining strong role consistency and interactive ability. The dataset and code are publicly available in https://github.com/Collab-Gen/DynSess.</abstract>
<identifier type="citekey">zhang-etal-2026-dynsess</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.689/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>14443</start>
<end>14461</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T DynSess: Dynamic Session-Level Evaluation and Optimization Framework for Role-Playing Agents
%A Zhang, Rongsheng
%A Tang, Jiji
%A Ren, Junnan
%A Bao, Zuyi
%A Chen, Weijie
%A Hu, Ruofan
%A Lv, Tangjie
%A Zhao, Zhou
%A Zhang, Yan
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F zhang-etal-2026-dynsess
%X Role-playing with large language models is fundamentally a *session-level* task, requiring agents to sustain character identity and interaction quality across extended multi-turn conversations. Yet existing evaluation and optimization methods remain largely turn-level, failing to capture long-horizon quality. We propose **DynSess**, a unified session-level framework for role-playing agents. **DynSess-Eval** scores complete dialogue sessions via rubrics targeting long-horizon behaviors. Leveraging its session-level rewards, we construct high-quality training trajectories through multi-turn lookahead search and train **DynSess-Character** with two complementary variants: DSPO (off-policy) and GSRPO (on-policy). Experiments show that DynSess-Eval aligns with human judgments substantially better than prior evaluators, and blind human evaluation further shows that DynSess-Character matches the strongest character model despite using substantially fewer parameters, while maintaining strong role consistency and interactive ability. The dataset and code are publicly available in https://github.com/Collab-Gen/DynSess.
%U https://aclanthology.org/2026.findings-emnlp.689/
%P 14443-14461
Markdown (Informal)
[DynSess: Dynamic Session-Level Evaluation and Optimization Framework for Role-Playing Agents](https://aclanthology.org/2026.findings-emnlp.689/) (Zhang et al., Findings 2026)
ACL
- Rongsheng Zhang, Jiji Tang, Junnan Ren, Zuyi Bao, Weijie Chen, Ruofan Hu, Tangjie Lv, Zhou Zhao, and Yan Zhang. 2026. DynSess: Dynamic Session-Level Evaluation and Optimization Framework for Role-Playing Agents. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 14443–14461, Budapest, Hungary. Association for Computational Linguistics.