On Inter-dataset Code Duplication and Data Leakage in Large Language Models

TL;DR

This paper investigates how code duplication across datasets can inflate the perceived performance of large language models in software engineering tasks, highlighting a potential evaluation bias.

Contribution

It empirically studies inter-dataset code duplication, demonstrating its impact on LLM evaluation and revealing vulnerabilities in current fine-tuning practices.

Findings

Inter-dataset code duplication can inflate LLM performance metrics.

Fine-tuning techniques influence the extent of evaluation bias.

Open-source models are also susceptible to data leakage from dataset overlaps.

Abstract

Motivation. Large language models (LLMs) have exhibited remarkable proficiency in diverse software engineering (SE) tasks. Handling such tasks typically involves acquiring foundational coding knowledge on large, general-purpose datasets during a pre-training phase, and subsequently refining on smaller, task-specific datasets as part of a fine-tuning phase. Problem statement. While intra-dataset code duplication examines the intersection between the training and test splits within a given dataset and has been addressed in prior research, inter-dataset code duplication, which gauges the overlap between different datasets, remains largely unexplored. If this phenomenon exists, it could compromise the integrity of LLM evaluations because of the inclusion of fine-tuning test samples that were already encountered during pre-training, resulting in inflated performance metrics. Contribution. This paper explores the phenomenon of inter-dataset code duplication and its impact on evaluating LLMs across diverse SE tasks. Study design. We conduct an empirical study using the CodeSearchNet dataset (CSN), a widely adopted pre-training dataset, and five fine-tuning datasets used for various se tasks. We first identify the intersection between the pre-training and fine-tuning datasets using a deduplication process. Next, we pre-train two versions of LLMs using a subset of CSN: one leaky LLM and one non-leaky LLM. Finally, we fine-tune both models and compare their performances using leaky fine-tuning test samples. Results. Our findings reveal a potential threat to the evaluation of LLMs across multiple SE tasks, stemming from the inter-dataset code duplication phenomenon. We also demonstrate that this threat is accentuated by the chosen fine-tuning technique. Furthermore, we provide evidence that open-source models could be affected by inter-dataset duplication.