Abstract:Empathetic response generation is a desirable aspect of conversational agents, crucial for facilitating engaging and emotionally intelligent multi-turn conversations between humans and machines. Leveraging large language models for this task has shown promising results, yet challenges persist in ensuring both the empathetic quality of the responses and retention of the generalization performance of the models. We propose a novel approach where we construct theory-driven preference datasets based on emotion grounding and use them to align LLMs with preference optimization algorithms to address these challenges. To evaluate empathetic response generation, we employ the EmpatheticDialogues dataset, assessing empathy with the diff-Epitome and BERTscore metrics and with multi-dimensional human evaluation. Additionally, we measure diversity and emotional valence using feature-based methods. We also evaluate the impact of training on the generalization performance using the MMLU benchmark and tasks from the Open LLM Leaderboard. The results show that LLMs can be aligned for empathetic response generation by preference optimization while retaining their general performance and that emotion grounding can guide preference dataset creation. We make all datasets, source code, and models publicly available. <a class="link-external link-https" href="https://github.com/justtherightsize/empo" rel="external noopener nofollow">this https URL</a>

Feel You ’ : Enhancing conversational agents with empathy Name :

Harnessing the Power of Large Language Models for Empathetic Response Generation: Empirical Investigations and Improvements

The Illusion of Empathy: How AI Chatbots Shape Conversation Perception

Exploring ChatGPT's Empathic Abilities

Multi-dimensional Evaluation of Empathetic Dialog Responses

Is ChatGPT More Empathetic than Humans?

CheerBots: Chatbots toward Empathy and Emotionusing Reinforcement Learning

Is ChatGPT Equipped with Emotional Dialogue Capabilities?

CAB: Empathetic Dialogue Generation with Cognition, Affection and Behavior

Towards Persona-Based Empathetic Conversational Models

Emotion-Aware Transformer Encoder for Empathetic Dialogue Generation

Empathetic Dialogue Generation with Pre-trained RoBERTa-GPT2 and External Knowledge

EmPO: Emotion Grounding for Empathetic Response Generation through Preference Optimization

EmpathyEar: An Open-source Avatar Multimodal Empathetic Chatbot

Harnessing Large Language Models' Empathetic Response Generation Capabilities for Online Mental Health Counselling Support

Using Adaptive Empathetic Responses for Teaching English

Endowing Empathetic Dialogue Systems with Personas

Wish I Can Feel What You Feel: A Neural Approach for Empathetic Response Generation

Assessing Empathy in Large Language Models with Real-World Physician-Patient Interactions

Empathetic Dialog Generation with Fine-Grained Intents

Use of a Taxonomy of Empathetic Response Intents to Control and Interpret Empathy in Neural Chatbots