Transaction

jmetaB025bcccdc76a45d416a7b353fe661ce6345c18debfc7282cfe7e02209766fe86e7@104e04f4dc7bbb58b675a0be8ec8a2392cd828cadc0c1b85347e2d4ab003150erss.itemmetarss.netM„<item> <title>ObfuscaTune: Obfuscated Offsite Fine-tuning and Inference of Proprietary LLMs on Private Datasets</title> <link>https://arxiv.org/abs/2407.02960</link> <description>arXiv:2407.02960v1 Announce Type: new Abstract: This work addresses the timely yet underexplored problem of performing inference and finetuning of a proprietary LLM owned by a model provider entity on the confidential/private data of another data owner entity, in a way that ensures the confidentiality of both the model and the data. Hereby, the finetuning is conducted offsite, i.e., on the computation infrastructure of a third-party cloud provider. We tackle this problem by proposing ObfuscaTune, a novel, efficient and fully utility-preserving approach that combines a simple yet effective obfuscation technique with an efficient usage of confidential computing (only 5% of the model parameters are placed on TEE). We empirically demonstrate the effectiveness of ObfuscaTune by validating it on GPT-2 models with different sizes on four NLP benchmark datasets. Finally, we compare to a na\"ive version of our approach to highlight the necessity of using random matrices with low condition numbers in our approach to reduce errors induced by the obfuscation.</description> <guid isPermaLink="false">oai:arXiv.org:2407.02960v1</guid> <category>cs.CR</category> <category>cs.AI</category> <category>cs.CL</category> <category>cs.LG</category> <arxiv:announce_type>new</arxiv:announce_type> <dc:rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0/</dc:rights> <dc:creator>Ahmed Frikha, Nassim Walha, Ricardo Mendes, Krishna Kanth Nakka, Xue Jiang, Xuebing Zhou</dc:creator> </item>