Enabling Fine-Grained Operating Points for Black-Box LLMs
This addresses the need for fine-grained adjustment of decision-making behavior in black-box LLMs for applications with specific metric constraints, representing an incremental improvement over existing methods.
The paper tackled the problem of limited control over operating points in black-box LLMs for classification tasks, showing that standard techniques fail to improve granularity without trade-offs, and proposed efficient approaches that significantly increase operating points while achieving comparable or better performance across 11 datasets and 3 LLMs.
Black-box Large Language Models (LLMs) provide practical and accessible alternatives to other machine learning methods, as they require minimal labeled data and machine learning expertise to develop solutions for various decision making problems. However, for applications that need operating with constraints on specific metrics (e.g., precision $\geq$ 95%), decision making with black-box LLMs remains unfavorable, due to their low numerical output cardinalities. This results in limited control over their operating points, preventing fine-grained adjustment of their decision making behavior. In this paper, we study using black-box LLMs as classifiers, focusing on efficiently improving their operational granularity without performance loss. Specifically, we first investigate the reasons behind their low-cardinality numerical outputs and show that they are biased towards generating rounded but informative verbalized probabilities. Then, we experiment with standard prompt engineering, uncertainty estimation and confidence elicitation techniques, and observe that they do not effectively improve operational granularity without sacrificing performance or increasing inference cost. Finally, we propose efficient approaches to significantly increase the number and diversity of available operating points. Our proposed approaches provide finer-grained operating points and achieve comparable to or better performance than the benchmark methods across 11 datasets and 3 LLMs.