Zero-to-Strong Generalization: Eliciting Strong Capabilities of Large Language Models Iteratively without Gold Labels

TL;DR

This paper introduces a novel zero-to-strong generalization paradigm that iteratively prompts LLMs to annotate unlabeled data, effectively unlocking their capabilities without relying on gold labels, and demonstrating success across diverse tasks and model sizes.

Contribution

The study proposes an iterative prompting framework that enables LLMs to improve their performance on tasks using only unlabeled data, bypassing the need for gold labels.

Findings

Effective on classification and reasoning tasks

Works for various model sizes and learning paradigms

Enhances LLM capabilities without gold labels

Abstract

Large Language Models (LLMs) have demonstrated remarkable performance through supervised fine-tuning or in-context learning using gold labels. However, this paradigm is limited by the availability of gold labels, while in certain scenarios, LLMs may need to perform tasks that are too complex for humans to provide such labels. To tackle this challenge, this study explores whether solely utilizing unlabeled data can elicit strong model capabilities. We propose a new paradigm termed zero-to-strong generalization. We iteratively prompt LLMs to annotate unlabeled data and retain high-quality labels by filtering. Surprisingly, we obverse that this iterative process gradually unlocks LLMs' potential on downstream tasks. Our experiments on extensive classification and reasoning tasks confirm the effectiveness of our proposed framework. Our analysis indicates that this paradigm is effective for both in-context learning and fine-tuning, and for various model sizes.