Generating Place-Based Compromises Between Two Points of View

2026-04-27

AI Summary

Large language models excel in academic tasks but struggle to balance opposing viewpoints in compromise-generation scenarios requiring social intelligence. This work proposes a novel paradigm centered on empathic neutrality, replacing conventional chain-of-thought reasoning with external empathic similarity as iterative feedback to achieve effective alignment without explicit empathic inference. By integrating prompt engineering, empathic similarity feedback, and margin-based human preference optimization, the approach significantly enhances the acceptability of generated compromises, as demonstrated in a user study with 50 participants. The methodology further enables the successful training of two lightweight foundation models, offering an efficient and scalable solution for socially nuanced negotiation tasks.

Technology Category

• Natural Language Processing: (Large) Language Models
• Humans and AI: Learning Human Values and Preferences
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Large Language Models (LLMs) excel academically but struggle with social intelligence tasks, such as creating good compromises. In this paper, we present methods for generating empathically neutral compromises between two opposing viewpoints. We first compared four different prompt engineering methods using Claude 3 Opus and a dataset of 2,400 contrasting views on shared places. A subset of the gen erated compromises was evaluated for acceptability in a 50-participant study. We found that the best method for generating compromises between two views used external empathic similarity between a compromise and each viewpoint as iterative feedback, outperforming stan dard Chain of Thought (CoT) reasoning. The results indicate that the use of empathic neutrality improves the acceptability of compromises. The dataset of generated compromises was then used to train two smaller foundation models via margin-based alignment of human preferences, improving efficiency and removing the need for empathy estimation during inference.

Problem

• compromise generation
• social intelligence
• empathic neutrality
• opposing viewpoints
• large language models

Innovation

• empathic neutrality
• compromise generation
• prompt engineering
• preference alignment
• large language models