Best Arm Identification for Prompt Learning under a Limited Budget
Keywords: Prompt Learning; Best-arm Identification; Limited Budget; Large Language Models
TL;DR: This work introduces a framework that harnesses the power of fixed-budget best arm identification in prompt learning for large language models (LLMs) under a limited budget, which achieves outstanding performance across tasks and targeted LLMs.
Abstract: The remarkable instruction-following capability of large language models (LLMs) has sparked a growing interest in automatically learning suitable prompts. However, while many effective methods have been proposed, the cost incurred during the learning process (e.g., accessing LLM and evaluating the responses) has not been considered. To overcome this limitation, this work explicitly incorporates a finite budget constraint into prompt learning. Towards developing principled solutions, a novel connection is established between prompt learning and fixed-budget best arm identification (BAI-FB) in multi-armed bandits (MAB). Based on this connection, a general framework TRIPLE (besT aRm Identification for Prompt LEarning) is proposed to harness the power of BAI-FB in prompt learning systematically. Extensive experiments on multiple well-adopted tasks using both GPT 3.5 and Llama2 demonstrate the superiority of TRIPLE over the previous baselines.
Submission Number: 72
Loading