Skip to yearly menu bar Skip to main content


Poster

InstructZero: Efficient Instruction Optimization for Black-Box Large Language Models

Lichang Chen · Jiuhai Chen · Tom Goldstein · Heng Huang · Tianyi Zhou

Hall C 4-9 #2909
[ ] [ Paper PDF ]
Tue 23 Jul 2:30 a.m. PDT — 4 a.m. PDT

Abstract:

Large language models (LLMs) are instruction followers but the performance varies under different instructions. It is challenging to create the best instruction, especially for black-box LLMs on which backpropagation is forbidden. Instead of directly optimizing the discrete instruction, we optimize a low-dimensional soft prompt applied to an open-source LLM to generate the instruction for the black-box LLM. In each optimization step of the proposed method InstructZero, a soft prompt is converted into an instruction by the open-source LLM, which is then submitted to the black-box LLM for zero-shot evaluation, whose result is sent to Bayesian optimization to produce new soft prompts improving the zero-shot performance. We evaluate InstructZero on different combinations of open-source LLMs and APIs including Vicuna and ChatGPT. InstructZero outperforms SOTA auto-instruction methods across a variety of downstream tasks.

Chat is not available.