Search papers, labs, and topics across Lattice.
The paper introduces CultureManager, a pipeline designed to align LLMs with specific cultural norms for culturally sensitive tasks by generating task-aware cultural data grounded in web search results. It addresses the limitations of existing cultural alignment methods that struggle with task specificity and cross-culture interference. The key result is consistent performance improvements across ten national cultures and culture-sensitive tasks compared to prompt-based and fine-tuning baselines, demonstrating the importance of task adaptation and modular culture management.
LLMs can be steered to respect cultural nuances in specific tasks by training modular adapters and routing between them, significantly improving performance over standard alignment techniques.
Large language models (LLMs) are increasingly deployed in culturally sensitive real-world tasks. However, existing cultural alignment approaches fail to align LLMs' broad cultural values with the specific goals of downstream tasks and suffer from cross-culture interference. We propose CultureManager, a novel pipeline for task-specific cultural alignment. CultureManager synthesizes task-aware cultural data in line with target task formats, grounded in culturally relevant web search results. To prevent conflicts between cultural norms, it manages multi-culture knowledge learned in separate adapters with a culture router that selects the appropriate one to apply. Experiments across ten national cultures and culture-sensitive tasks show consistent improvements over prompt-based and fine-tuning baselines. Our results demonstrate the necessity of task adaptation and modular culture management for effective cultural alignment.
