Saved in:
Bibliographic Details
Main Authors: Wang, Weixing, Yang, Haojin, Meinel, Christoph
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2406.14208
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917787263303680
author Wang, Weixing
Yang, Haojin
Meinel, Christoph
author_facet Wang, Weixing
Yang, Haojin
Meinel, Christoph
contents Previous studies have shown that demonstrations can significantly help Large Language Models (LLMs ) perform better on the given tasks. However, this so-called In-Context Learning ( ICL ) ability is very sensitive to the presenting context, and often dozens of demonstrations are needed. In this work, we investigate if we can reduce the shot number while still maintaining a competitive performance. We present SeCoKD, a self-Knowledge Distillation ( KD ) training framework that aligns the student model with a heavily prompted variation, thereby increasing the utilization of a single demonstration. We experiment with the SeCoKD across three LLMs and six benchmarks focusing mainly on reasoning tasks. Results show that our method outperforms the base model and Supervised Fine-tuning ( SFT ), especially in zero-shot and one-shot settings by 30% and 10%, respectively. Moreover, SeCoKD brings little negative artifacts when evaluated on new tasks, which is more robust than Supervised Fine-tuning.
format Preprint
id arxiv_https___arxiv_org_abs_2406_14208
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SeCoKD: Aligning Large Language Models for In-Context Learning with Fewer Shots
Wang, Weixing
Yang, Haojin
Meinel, Christoph
Artificial Intelligence
Previous studies have shown that demonstrations can significantly help Large Language Models (LLMs ) perform better on the given tasks. However, this so-called In-Context Learning ( ICL ) ability is very sensitive to the presenting context, and often dozens of demonstrations are needed. In this work, we investigate if we can reduce the shot number while still maintaining a competitive performance. We present SeCoKD, a self-Knowledge Distillation ( KD ) training framework that aligns the student model with a heavily prompted variation, thereby increasing the utilization of a single demonstration. We experiment with the SeCoKD across three LLMs and six benchmarks focusing mainly on reasoning tasks. Results show that our method outperforms the base model and Supervised Fine-tuning ( SFT ), especially in zero-shot and one-shot settings by 30% and 10%, respectively. Moreover, SeCoKD brings little negative artifacts when evaluated on new tasks, which is more robust than Supervised Fine-tuning.
title SeCoKD: Aligning Large Language Models for In-Context Learning with Fewer Shots
topic Artificial Intelligence
url https://arxiv.org/abs/2406.14208