Research map / Large language models / Agents, code and math
R4: Nested Reasoning-Retrieval for Reward Modeling in Role-Playing Agents
Renzhi Wang, Chongqiang Wei, Zhisheng Wang, Piji Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add reasoning to reward model and agent (ablation) | R4 · CharacterEval Character Consistency | average score: 48.23 -> 64.64 | Improved |
| add retrieval to reward model and agent (ablation) | R4 · CharacterEval Character Consistency | average score: 51.67 -> 64.64 | Improved |
| use character-specific instead of generic retrieval (ablation) | R4 · CharacterEval Character Consistency | average score: 47.95 -> 64.64 | Improved |
| add reasoning to the reward model (ablation) | R4 · CharacterEval Character Consistency | average score: 49.87 -> 64.64 | Improved |
| add retrieval to the reward model (ablation) | R4 · CharacterEval Character Consistency | average score: 58.31 -> 64.64 | Improved |
| add reasoning to the agent (ablation) | R4 · CharacterEval Character Consistency | average score: 56.94 -> 64.64 | Improved |
| add retrieval to the agent (ablation) | R4 · CharacterEval Character Consistency | average score: 55.47 -> 64.64 | Improved |
| add consistency reward to RM (ablation) | R4 · CharacterEval Character Consistency | average score: 59.83 -> 64.64 | Improved |
| add format reward to RM (ablation) | R4 · CharacterEval Character Consistency | average score: 62.41 -> 64.64 | Improved |
| use R4 reasoning-retrieval RM instead of scalar RM (ablation) | R4 · CharacterEval Character Consistency | average score: 43.89 -> 64.64 | Improved |
| use R4 RM instead of CharacterRM (ablation) | R4 · CharacterEval Character Consistency | average score: 46.87 -> 64.64 | Improved |
| scale agent from 7B to 32B | R4 · CharacterEval Character Consistency | average score: 55.74 -> 64.64 | Improved |
Closest papers by meaning
- Thinking in Character: Advancing Role-Playing Agents with Role-Aware Reasoning (NeurIPS 2025 · similarity 0.90)
- RLVER: Reinforcement Learning with Verifiable Emotion Rewards for Empathetic Agents (ICLR 2026 · similarity 0.89)
- Consistently Simulating Human Personas with Multi-Turn Reinforcement Learning (NeurIPS 2025 · similarity 0.88)
- Codifying Character Logic in Role-Playing (NeurIPS 2025 · similarity 0.87)
- Memory-T1: Reinforcement Learning for Temporal Reasoning in Multi-session Agents (ICLR 2026 · similarity 0.87)
- Concept Incongruence: An Exploration of Time and Death in Role Playing (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
