The recent interest in the deployment of Generative AI applications that use large language models (LLMs) has brought to the forefront significant privacy concerns, notably the leakage of Personally Identifiable Infor...The recent interest in the deployment of Generative AI applications that use large language models (LLMs) has brought to the forefront significant privacy concerns, notably the leakage of Personally Identifiable Information (PII) and other confidential or protected information that may have been memorized during training, specifically during a fine-tuning or customization process. We describe different black-box attacks from potential adversaries and study their impact on the amount and type of information that may be recovered from commonly used and deployed LLMs. Our research investigates the relationship between PII leakage, memorization, and factors such as model size, architecture, and the nature of attacks employed. The study utilizes two broad categories of attacks: PII leakage-focused attacks (auto-completion and extraction attacks) and memorization-focused attacks (various membership inference attacks). The findings from these investigations are quantified using an array of evaluative metrics, providing a detailed understanding of LLM vulnerabilities and the effectiveness of different attacks.展开更多
The recent interest in the deployment of Generative AI applications that use large language models (LLMs) has brought to the forefront significant privacy concerns, notably the leakage of Personally Identifiable Infor...The recent interest in the deployment of Generative AI applications that use large language models (LLMs) has brought to the forefront significant privacy concerns, notably the leakage of Personally Identifiable Information (PII) and other confidential or protected information that may have been memorized during training, specifically during a fine-tuning or customization process. This inadvertent leakage of sensitive information typically occurs when the models are subjected to black-box attacks. To address the growing concerns of safeguarding private and sensitive information while simultaneously preserving its utility, we analyze the performance of Targeted Catastrophic Forgetting (TCF). TCF involves preserving targeted pieces of sensitive information within datasets through an iterative pipeline which significantly reduces the likelihood of such information being leaked or reproduced by the model during black-box attacks, such as the autocompletion attack in our case. The experiments conducted using TCF evidently demonstrate its capability to reduce the extraction of PII while still preserving the context and utility of the target application.展开更多
Rauwolfia species(Apocynaceae) are medicinal plants well known worldwide due to its potent bioactive monoterpene indole alkaloids(MIAs) such as reserpine,ajmalicine,ajmaline,serpentine and yohimbine.Reserpine,ajmalici...Rauwolfia species(Apocynaceae) are medicinal plants well known worldwide due to its potent bioactive monoterpene indole alkaloids(MIAs) such as reserpine,ajmalicine,ajmaline,serpentine and yohimbine.Reserpine,ajmalicine and ajmaline are powerful antihypertensive,tranquilizing agents used in hypertension.Yohimbine is an aphrodisiac used in dietary supplements.As there is no report on the comparative and comprehensive phytochemical investigation of the roots of Rauwolfia species,we have developed an efficient and reliable liquid chromatography-tandem mass spectrometry(LC–MS/MS) method for ethanolic root extract of Rauwolfia species to elucidate the fragmentation pathways for dereplication of bioactive MIAs using highperformance liquid chromatography coupled with electrospray ionization quadrupole time-of-flight tandem mass spectrometry(HPLC–ESI–QTOF–MS/MS) in positive ion mode.We identified and established diagnostic fragment ions and fragmentation pathways using reserpine,ajmalicine,ajmaline,serpentine and yohimbine.The MS/MS spectra of reserpine,ajmalicine,and ajmaline showed C-ring-cleavage whereas E-ring cleavage was observed in serpentine via Retro Diels Alder(RDA).A total of 47 bioactive MIAs were identified and characterized on the basis of their molecular formula,exact mass measurements and MS/MS analysis.Reserpine,ajmalicine,ajmaline,serpentine and yohimbine were unambiguously identified by comparison with their authentic standards and other 42 MIAs were tentatively identified and characterized from the roots of Rauwolfia hookeri,Rauwolfia micrantha,Rauwolfia serpentina,Rauwolfia verticillata,Rauwolfia tetraphylla and Rauwolfia vomitoria.Application of LC–MS followed by principal component analysis(PCA) has been successfully used to discriminate among six Rauwolfia species.展开更多
文摘The recent interest in the deployment of Generative AI applications that use large language models (LLMs) has brought to the forefront significant privacy concerns, notably the leakage of Personally Identifiable Information (PII) and other confidential or protected information that may have been memorized during training, specifically during a fine-tuning or customization process. We describe different black-box attacks from potential adversaries and study their impact on the amount and type of information that may be recovered from commonly used and deployed LLMs. Our research investigates the relationship between PII leakage, memorization, and factors such as model size, architecture, and the nature of attacks employed. The study utilizes two broad categories of attacks: PII leakage-focused attacks (auto-completion and extraction attacks) and memorization-focused attacks (various membership inference attacks). The findings from these investigations are quantified using an array of evaluative metrics, providing a detailed understanding of LLM vulnerabilities and the effectiveness of different attacks.
文摘The recent interest in the deployment of Generative AI applications that use large language models (LLMs) has brought to the forefront significant privacy concerns, notably the leakage of Personally Identifiable Information (PII) and other confidential or protected information that may have been memorized during training, specifically during a fine-tuning or customization process. This inadvertent leakage of sensitive information typically occurs when the models are subjected to black-box attacks. To address the growing concerns of safeguarding private and sensitive information while simultaneously preserving its utility, we analyze the performance of Targeted Catastrophic Forgetting (TCF). TCF involves preserving targeted pieces of sensitive information within datasets through an iterative pipeline which significantly reduces the likelihood of such information being leaked or reproduced by the model during black-box attacks, such as the autocompletion attack in our case. The experiments conducted using TCF evidently demonstrate its capability to reduce the extraction of PII while still preserving the context and utility of the target application.
基金Council of Scientific Industrial Research,India for providing financial support
文摘Rauwolfia species(Apocynaceae) are medicinal plants well known worldwide due to its potent bioactive monoterpene indole alkaloids(MIAs) such as reserpine,ajmalicine,ajmaline,serpentine and yohimbine.Reserpine,ajmalicine and ajmaline are powerful antihypertensive,tranquilizing agents used in hypertension.Yohimbine is an aphrodisiac used in dietary supplements.As there is no report on the comparative and comprehensive phytochemical investigation of the roots of Rauwolfia species,we have developed an efficient and reliable liquid chromatography-tandem mass spectrometry(LC–MS/MS) method for ethanolic root extract of Rauwolfia species to elucidate the fragmentation pathways for dereplication of bioactive MIAs using highperformance liquid chromatography coupled with electrospray ionization quadrupole time-of-flight tandem mass spectrometry(HPLC–ESI–QTOF–MS/MS) in positive ion mode.We identified and established diagnostic fragment ions and fragmentation pathways using reserpine,ajmalicine,ajmaline,serpentine and yohimbine.The MS/MS spectra of reserpine,ajmalicine,and ajmaline showed C-ring-cleavage whereas E-ring cleavage was observed in serpentine via Retro Diels Alder(RDA).A total of 47 bioactive MIAs were identified and characterized on the basis of their molecular formula,exact mass measurements and MS/MS analysis.Reserpine,ajmalicine,ajmaline,serpentine and yohimbine were unambiguously identified by comparison with their authentic standards and other 42 MIAs were tentatively identified and characterized from the roots of Rauwolfia hookeri,Rauwolfia micrantha,Rauwolfia serpentina,Rauwolfia verticillata,Rauwolfia tetraphylla and Rauwolfia vomitoria.Application of LC–MS followed by principal component analysis(PCA) has been successfully used to discriminate among six Rauwolfia species.
