Detecting Data Contamination from Reinforcement Learning Post-training for Large Language Models

ICLR 2026 Conference Submission12983 Authors

Keywords: Data Contamination Detection; LLMs; Reinforcement Learning; Entropy

TL;DR: We present the first study of data contamination detection in RL post-training, introducing the RL-MIA benchmark and an entropy-based method Self-Critique, which achieving an AUC improvement of up to 30% over prior baselines.

Abstract:

Data contamination poses a significant threat to the reliable evaluation of Large Language Models (LLMs). This issue arises when benchmark samples may inadvertently appear in training sets, compromising the validity of reported performance. While detection methods have been developed for the pre-training and Supervised Fine-Tuning stages, a critical research gap exists for the increasingly significant phase of Reinforcement Learning (RL) post-training.
As RL post-training becomes pivotal for advancing LLM reasoning, the absence of specialized contamination detection methods in this paradigm presents a critical vulnerability.
To address this, we conduct the first systematic study of data detection within RL post-training scenario and propose Self-Critique. Our method is motivated by a key observation: after RL phase, the output entropy distribution of LLMs tends to collapse into highly specific and sparse modes. Self-Critique probes for the underlying policy collapse, i.e., the model's convergence to a narrow reasoning path, which causes this entropy reduction.
To facilitate this research, we also introduce RL-MIA, a benchmark constructed to simulate this specific contamination scenario.
Extensive experiments show that Self-Critique significantly outperforms baseline methods across multiple models and contamination tasks, achieving an AUC improvement of up to 30%. Whereas existing methods are close to a random guess for RL-phase contamination, our method makes detection possible.
Our benchmark and code are available at https://anonymous.4open.science/r/Data-Contamination-in-RL.

Primary Area: alignment, fairness, safety, privacy, and societal considerations

Submission Number: 12983