Flatness-Aware Prompt Selection Improves Accuracy and Sample Efficiency
This work addresses the challenge of prompt selection for users of large language models, offering an incremental improvement over existing methods.
The paper tackles the problem of automatically selecting effective prompts for large language models by introducing a new metric called prompt flatness, which quantifies a prompt's expected utility based on robustness to parameter perturbations. The result shows that combining this metric with existing ones improves accuracy by an average of 5% and Pearson correlation by 10% across six classification benchmarks.
With growing capabilities of large language models, prompting them has become the dominant way to access them. This has motivated the development of strategies for automatically selecting effective language prompts. In this paper, we introduce prompt flatness, a new metric to quantify the expected utility of a language prompt. This metric is inspired by flatness regularization in statistical learning that quantifies the robustness of the model towards its parameter perturbations. We provide theoretical foundations for this metric and its relationship with other prompt selection metrics, providing a comprehensive understanding of existing methods. Empirically, we show that combining prompt flatness with existing metrics improves both performance and sample efficiency. Our metric outperforms the previous prompt selection metrics with an average increase of 5% in accuracy and 10% in Pearson correlation across 6 classification benchmarks.