G-Boost: Boosting Private SLMs with General LLMs
This addresses the challenge for developers with limited computational resources who rely on private SLMs, offering an incremental improvement by leveraging existing general LLMs.
The paper tackles the problem of limited effectiveness in private small language models (SLMs) by proposing the G-Boost framework, which enables adaptive collaborative inference with general LLMs, resulting in significant performance boosts as demonstrated in experiments.
Due to the limited computational resources, most Large Language Models (LLMs) developers can only fine-tune Small Language Models (SLMs) on their own data. These private SLMs typically have limited effectiveness. To boost the performance of private SLMs, this paper proposes to ask general LLMs for help. The general LLMs can be APIs or larger LLMs whose inference cost the developers can afford. Specifically, we propose the G-Boost framework where a private SLM adaptively performs collaborative inference with a general LLM under the guide of process reward. Experiments demonstrate that our framework can significantly boost the performance of private SLMs.