咨询下，LLM的数据污染检测（判断数据集是否训练见过）技术方向靠谱吗？有推荐论文吗？

[gongjunjin]

[cyp-jlu-ai]

No description provided.

Hello, data contamination detection for LLM is an important research area, especially in ensuring the quality and reliability of model training data.

Here are some recommended papers on data contamination detection for LLM:

1. Time Travel in LLMs: Tracing Data Contamination in Large Language Models
2. Can we trust the evaluation on ChatGPT?
3. Large language models are few-shot testers: Exploring llm-based general bug reproduction
4. BAMBOO: A Comprehensive Benchmark for Evaluating Long Text Modeling Capacities of Large Language Models
5. Anatomy of an AI-powered malicious social botnet
6. Origin Tracing and Detecting of LLMs