Modular Pluralism: Pluralistic Alignment via Multi-LLM Collaboration
This addresses the need for AI systems to better represent diverse communities, though it appears incremental as it builds on existing alignment paradigms with a modular approach.
The paper tackles the problem of large language models struggling to model diverse human preferences across cultures and demographics by proposing Modular Pluralism, a framework that uses multi-LLM collaboration to support pluralistic alignment, with experiments showing advancements in three pluralism objectives across six tasks and datasets.
While existing alignment paradigms have been integral in developing large language models (LLMs), LLMs often learn an averaged human preference and struggle to model diverse preferences across cultures, demographics, and communities. We propose Modular Pluralism, a modular framework based on multi-LLM collaboration for pluralistic alignment: it "plugs into" a base LLM a pool of smaller but specialized community LMs, where models collaborate in distinct modes to flexibility support three modes of pluralism: Overton, steerable, and distributional. Modular Pluralism is uniquely compatible with black-box LLMs and offers the modular control of adding new community LMs for previously underrepresented communities. We evaluate Modular Pluralism with six tasks and four datasets featuring questions/instructions with value-laden and perspective-informed responses. Extensive experiments demonstrate that Modular Pluralism advances the three pluralism objectives across six black-box and open-source LLMs. Further analysis reveals that LLMs are generally faithful to the inputs from smaller community LLMs, allowing seamless patching by adding a new community LM to better cover previously underrepresented communities.