mlsanityData to 14 Oct 2026

DynSess: Dynamic Session-Level Evaluation and Optimization Framework for Role-Playing Agents

arXiv 2605.29256 ↗First posted 28 May 2026cs.CL cs.AIcited by 2

On the map: LLM agents and tool use → Social simulation and multi-agent debate

Abstract (arXiv)

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. Our dataset and code will be released to facilitate future research.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use. Builds on 20 of its 34 references that are in the map. First posted on arXiv 28 May 2026.

Builds on

20 of 34 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024CharacterEval: A Chinese Benchmark for Role-Playing Conversational Agent Evaluation2 Jan

Other references · 19

  1. 2024LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Modelscited by 1,464
  2. 2024ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RLcited by 228
  3. 2024Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalizationcited by 208
  4. 2023InCharacter: Evaluating Personality Fidelity in Role-Playing Agents through Psychological Interviewscited by 179
  5. 2025CollabLLM: From Passive Responders to Active Collaboratorscited by 124
  6. 2024Large Language Models are Superpositions of All Characters: Attaining Arbitrary Role-play via Self-Alignmentcited by 104
  7. 2024CheckEval: A reliable LLM-as-a-Judge framework for evaluating text generation using checklistscited by 94
  8. 2024Direct Multi-Turn Preference Optimization for Language Agentscited by 64
  9. 2023ChatHaruhi: Reviving Anime Character in Reality via Large Language Modelcited by 59
  10. 2025RocketEval: Efficient Automated LLM Evaluation via Grading Checklistcited by 39
  11. 2024Regressing the Relative Future: Efficient Policy Optimization for Multi-turn RLHFcited by 30
  12. 2025SDPO: Segment-Level Direct Preference Optimization for Social Agentscited by 30
7 more
  1. 2023Aligning Large Language Models with Human Opinions through Persona Selection and Value--Belief--Norm Reasoningcited by 19
  2. 2025TRACT: Regression-Aware Fine-tuning Meets Chain-of-Thought Reasoning for LLM-as-a-Judgecited by 17
  3. 2025Reasoning Does Not Necessarily Improve Role-Playing Abilitycited by 14
  4. 2025RMTBench: Benchmarking LLMs Through Multi-Turn User-Centric Role-Playingcited by 10
  5. 2025CPO: Addressing Reward Ambiguity in Role-playing Dialogue via Comparative Policy Optimizationcited by 10
  6. 2026HER: Human-like Reasoning and Reinforcement Learning for LLM Role-playingcited by 10
  7. 2025Single LLM, Multiple Roles: A Unified Retrieval-Augmented Generation Framework Using Role-Specific Token Optimizationcited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026TRACE Bench: Task-driven Roleplay Agentic Checklist Evaluationcited by 1
  2. 2026SocialCoach: Personalized Social Skill Learning with Agentic Tutoring and Practicecited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Test-Time-Matching: Decouple Personality, Memory, and Linguistic Style in LLM-based Role-Playing Language Agent
  2. 2025SpeechRole: A Large-Scale Dataset and Benchmark for Evaluating Speech Role-Playing Agents
  3. 2026RoleCDE:Benchmarking and Mitigating Role-Alignment Trade-offs in Role-Playing Agents
  4. 2025FURINA: A Fully Customizable Role-Playing Benchmark via Scalable Multi-Agent Collaboration Pipeline
  5. 2026Adversarial Stress Testing of Role-Playing Language Agents using Multi-Agent Evaluation
  6. 2026Adversarial Closed-Loop Curriculum for Evolving Role-Playing Agents
  7. 2024ERABAL: Enhancing Role-Playing Agents through Boundary-Aware Learning
  8. 2025RoleMRC: A Fine-Grained Composite Benchmark for Role-Playing and Instruction-Following