RLVER: Reinforcement Learning with Verifiable Emotion Rewards for Empathetic Agents
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910039753621504 |
|---|---|
| author | Wang, Peisong Ma, Ruotian Zhang, Bang Chen, Xingyu He, Zhiwei Luo, Kang Lv, Qingsong Jiang, Qingxuan Xie, Zheng Wang, Shanyi Li, Yuan Ye, Fanghua Li, Jian Yang, Yifan Tu, Zhaopeng Li, Xiaolong |
| author_facet | Wang, Peisong Ma, Ruotian Zhang, Bang Chen, Xingyu He, Zhiwei Luo, Kang Lv, Qingsong Jiang, Qingxuan Xie, Zheng Wang, Shanyi Li, Yuan Ye, Fanghua Li, Jian Yang, Yifan Tu, Zhaopeng Li, Xiaolong |
| contents | Large language models (LLMs) excel at logical and algorithmic reasoning, yet their emotional intelligence (EQ) still lags far behind their cognitive prowess. While reinforcement learning from verifiable rewards (RLVR) has advanced in other domains, its application to dialogue-especially for emotional intelligence-remains underexplored. In this work, we introduce RLVER, the first end-to-end reinforcement learning framework that leverages verifiable emotion rewards from simulated users to cultivate higher-order empathetic abilities in LLMs. Within this framework, self-consistent affective simulated users engage in dialogue rollouts and produce deterministic emotion scores during conversations, serving as reward signals to guide the LLM's learning. Fine-tuning publicly available Qwen2.5-7B-Instruct model with PPO boosts its Sentient-Benchmark score from 13.3 to 79.2 while largely preserving mathematical and coding competence. Extensive experiments reveal that: (i) RLVER consistently improves multiple dialogue capabilities; (ii) Thinking and non-thinking models show distinct trends--thinking models excel in empathy and insight, while non-thinking models favor action; (iii) GRPO often yields stable gains, while PPO can push certain capabilities to a higher ceiling; (iv) More challenging environments are not always better-moderate ones can yield stronger outcomes. Our results show that RLVER is a practical route toward emotionally intelligent and broadly capable language agents. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_03112 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RLVER: Reinforcement Learning with Verifiable Emotion Rewards for Empathetic Agents Wang, Peisong Ma, Ruotian Zhang, Bang Chen, Xingyu He, Zhiwei Luo, Kang Lv, Qingsong Jiang, Qingxuan Xie, Zheng Wang, Shanyi Li, Yuan Ye, Fanghua Li, Jian Yang, Yifan Tu, Zhaopeng Li, Xiaolong Computation and Language Artificial Intelligence Computers and Society Large language models (LLMs) excel at logical and algorithmic reasoning, yet their emotional intelligence (EQ) still lags far behind their cognitive prowess. While reinforcement learning from verifiable rewards (RLVR) has advanced in other domains, its application to dialogue-especially for emotional intelligence-remains underexplored. In this work, we introduce RLVER, the first end-to-end reinforcement learning framework that leverages verifiable emotion rewards from simulated users to cultivate higher-order empathetic abilities in LLMs. Within this framework, self-consistent affective simulated users engage in dialogue rollouts and produce deterministic emotion scores during conversations, serving as reward signals to guide the LLM's learning. Fine-tuning publicly available Qwen2.5-7B-Instruct model with PPO boosts its Sentient-Benchmark score from 13.3 to 79.2 while largely preserving mathematical and coding competence. Extensive experiments reveal that: (i) RLVER consistently improves multiple dialogue capabilities; (ii) Thinking and non-thinking models show distinct trends--thinking models excel in empathy and insight, while non-thinking models favor action; (iii) GRPO often yields stable gains, while PPO can push certain capabilities to a higher ceiling; (iv) More challenging environments are not always better-moderate ones can yield stronger outcomes. Our results show that RLVER is a practical route toward emotionally intelligent and broadly capable language agents. |
| title | RLVER: Reinforcement Learning with Verifiable Emotion Rewards for Empathetic Agents |
| topic | Computation and Language Artificial Intelligence Computers and Society |
| url | https://arxiv.org/abs/2507.03112 |