Instruction Tuning for Large Language Models: A Survey
Instruction tuning surveys supervised fine-tuning of LLMs on instruction-output pairs to align next-word prediction with human intent, covering datasets, training, applications, and limitations.
Published Nov 17, 202579 citationsPaper ↗
Only vote on papers you've read. Sign in with GitHub to vote.
A vital survey for cataloging instruction tuning's messy definitions and past failures, though it risks bibliography padding over isolating whether format alignment drives real reasoning gains.
Abstract
This article surveys research works in the quickly advancing field of instruction tuning (IT), a crucial technique to enhance the capabilities and controllability of large language models (LLMs). Instruction tuning refers to the process of further training LLMs on a dataset consisting of (instruction, output) pairs in a supervised fashion, which bridges the gap between the next-word prediction objective of LLMs and the users’ objective of having LLMs adhere to human instructions. In this work, we make a systematic review of the literature, including the general methodology of IT, the construction of IT datasets, the training of IT models, and applications to different modalities, domains and application, along with analysis of aspects that influence the outcome of IT (e.g., generation of instruction outputs, size of the instruction dataset). We also review the potential pitfalls of IT along with criticism against it, along with efforts pointing out current deficiencies of existing strategies and suggest some avenues for fruitful research.