BILLY: Steering Large Language Models via Merging Persona Vectors for Creative Generation
This addresses the problem of inefficient multi-LLM collaboration for creative tasks, offering a more practical solution for users in AI-driven creativity applications, though it is incremental as it builds on existing persona vector methods.
The paper tackles the high computational costs and latency of multi-LLM systems for creative generation by proposing BILLY, a training-free framework that blends persona vectors in a single model to achieve diverse perspectives, resulting in improved performance on creativity benchmarks and reduced inference time and costs.
Multi-LLM systems enhance the creativity of large language models by simulating human collective intelligence but suffer from significant drawbacks, such as high computational costs and inference latency. To address these limitations, we propose BILLY (BlendIng persona vectors for Large Language model creativitY), a training-free framework that captures the benefits of multi-LLM collaboration, i.e. inducing diverse perspectives and specialized expertise, within a single model. BILLY operates by extracting and blending multiple distinct persona vectors directly in the model's activation space. We steer the model's generation process with this merged vector while inference, enabling multi-perspective output without explicit multi-LLM communication. Our experiments across creativity-oriented benchmarks demonstrate that BILLY surpasses single model prompting and traditional multi-LLM approaches, while substantially reducing inference time and computational costs. Our analyses further reveal that distinct persona vectors can be blended to achieve both effective control over complementary aspects of generation and greater interpretability.