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

arXiv: https://arxiv.org/abs/2605.29256 · HTML: https://mlsanity.com/paper/2605.29256/

## 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.

## Placed in

- LLM agents and tool use / [Social simulation and multi-agent debate](https://mlsanity.com/agents/social-simulation-and-multi-agent-debate/)

## Builds on (most cited first, 20 of 20)

- [LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models](https://mlsanity.com/paper/2403.13372/) (2024)
- [ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL](https://mlsanity.com/paper/2402.19446/) (2024)
- [Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization](https://mlsanity.com/paper/2406.01171/) (2024)
- [InCharacter: Evaluating Personality Fidelity in Role-Playing Agents through Psychological Interviews](https://mlsanity.com/paper/2310.17976/) (2023)
- [CharacterEval: A Chinese Benchmark for Role-Playing Conversational Agent Evaluation](https://mlsanity.com/paper/2401.01275/) (2024)
- [CollabLLM: From Passive Responders to Active Collaborators](https://mlsanity.com/paper/2502.00640/) (2025)
- [Large Language Models are Superpositions of All Characters: Attaining Arbitrary Role-play via Self-Alignment](https://mlsanity.com/paper/2401.12474/) (2024)
- [CheckEval: A reliable LLM-as-a-Judge framework for evaluating text generation using checklists](https://mlsanity.com/paper/2403.18771/) (2024)
- [Direct Multi-Turn Preference Optimization for Language Agents](https://mlsanity.com/paper/2406.14868/) (2024)
- [ChatHaruhi: Reviving Anime Character in Reality via Large Language Model](https://mlsanity.com/paper/2308.09597/) (2023)
- [RocketEval: Efficient Automated LLM Evaluation via Grading Checklist](https://mlsanity.com/paper/2503.05142/) (2025)
- [Regressing the Relative Future: Efficient Policy Optimization for Multi-turn RLHF](https://mlsanity.com/paper/2410.04612/) (2024)
- [SDPO: Segment-Level Direct Preference Optimization for Social Agents](https://mlsanity.com/paper/2501.01821/) (2025)
- [Aligning Large Language Models with Human Opinions through Persona Selection and Value--Belief--Norm Reasoning](https://mlsanity.com/paper/2311.08385/) (2023)
- [TRACT: Regression-Aware Fine-tuning Meets Chain-of-Thought Reasoning for LLM-as-a-Judge](https://mlsanity.com/paper/2503.04381/) (2025)
- [Reasoning Does Not Necessarily Improve Role-Playing Ability](https://mlsanity.com/paper/2502.16940/) (2025)
- [RMTBench: Benchmarking LLMs Through Multi-Turn User-Centric Role-Playing](https://mlsanity.com/paper/2507.20352/) (2025)
- [CPO: Addressing Reward Ambiguity in Role-playing Dialogue via Comparative Policy Optimization](https://mlsanity.com/paper/2508.09074/) (2025)
- [HER: Human-like Reasoning and Reinforcement Learning for LLM Role-playing](https://mlsanity.com/paper/2601.21459/) (2026)
- [Single LLM, Multiple Roles: A Unified Retrieval-Augmented Generation Framework Using Role-Specific Token Optimization](https://mlsanity.com/paper/2505.15444/) (2025)

## Built on by (most cited first, 2 of 2)

- [TRACE Bench: Task-driven Roleplay Agentic Checklist Evaluation](https://mlsanity.com/paper/2608.11236/) (2026)
- [SocialCoach: Personalized Social Skill Learning with Agentic Tutoring and Practice](https://mlsanity.com/paper/2606.04155/) (2026)

## 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.