ClusterLLM: Large Language Models as a Guide for Text Clustering

TL;DR

ClusterLLM introduces a novel framework that uses instruction-tuned large language models like ChatGPT to improve text clustering by leveraging feedback and user preferences, outperforming traditional methods.

Contribution

It presents a new approach that utilizes LLM feedback for clustering, enabling understanding of user preferences and improving clustering quality without access to embeddings.

Findings

Consistently improves clustering quality across 14 datasets.

Cost-effective with an average cost of ~$0.6 per dataset.

Effective for fine-tuning small embedders using ChatGPT feedback.

Abstract

We introduce ClusterLLM, a novel text clustering framework that leverages feedback from an instruction-tuned large language model, such as ChatGPT. Compared with traditional unsupervised methods that builds upon "small" embedders, ClusterLLM exhibits two intriguing advantages: (1) it enjoys the emergent capability of LLM even if its embeddings are inaccessible; and (2) it understands the user's preference on clustering through textual instruction and/or a few annotated data. First, we prompt ChatGPT for insights on clustering perspective by constructing hard triplet questions <does A better correspond to B than C>, where A, B and C are similar data points that belong to different clusters according to small embedder. We empirically show that this strategy is both effective for fine-tuning small embedder and cost-efficient to query ChatGPT. Second, we prompt ChatGPT for helps on clustering granularity by carefully designed pairwise questions <do A and B belong to the same category>, and tune the granularity from cluster hierarchies that is the most consistent with the ChatGPT answers. Extensive experiments on 14 datasets show that ClusterLLM consistently improves clustering quality, at an average cost of ~$0.6 per dataset. The code will be available at https://github.com/zhang-yu-wei/ClusterLLM.