Instructive Decoding: Instruction-Tuned Large Language Models are Self-Refiner from Noisy Instructions

arXiv:2311.00233v2 Announce Type: replace
Abstract: While instruction-tuned language models have demonstrated impressive zero-shot generalization, these models often struggle to generate accurate responses when faced with instructions that fall outside their training set. This paper presents Instructive Decoding (ID), a simple yet effective approach that augments the efficacy of instruction-tuned models. Specifically, ID adjusts the logits for next-token prediction in a contrastive manner, utilizing predictions generated from a manipulated version of the original instruction, referred to as a noisy instruction. This …

abstract arxiv cs.cl decoding generate instruction-tuned language language models large language large language models paper responses set simple struggle training type zero-shot