Survival of the Safest: Towards Secure Prompt Optimization through Interleaved Multi-Objective Evolution
Large language models (LLMs) have demonstrated remarkable capabilities; however, the optimization of their prompts has historically prioritized performance metrics at the expense of crucial safety and security considerations. To overcome this shortcoming, we introduce "Survival of the Safest&qu...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2024-10 |
---|---|
Hauptverfasser: | , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | arXiv.org |
container_volume | |
creator | Sinha, Ankita Cui, Wendi Das, Kamalika Zhang, Jiaxin |
description | Large language models (LLMs) have demonstrated remarkable capabilities; however, the optimization of their prompts has historically prioritized performance metrics at the expense of crucial safety and security considerations. To overcome this shortcoming, we introduce "Survival of the Safest" (SoS), an innovative multi-objective prompt optimization framework that enhances both performance and security in LLMs simultaneously. SoS utilizes an interleaved multi-objective evolution strategy, integrating semantic, feedback, and crossover mutations to effectively traverse the prompt landscape. Differing from the computationally demanding Pareto front methods, SoS provides a scalable solution that expedites optimization in complex, high-dimensional discrete search spaces while keeping computational demands low. Our approach accommodates flexible weighting of objectives and generates a pool of optimized candidates, empowering users to select prompts that optimally meet their specific performance and security needs. Experimental evaluations across diverse benchmark datasets affirm SoS's efficacy in delivering high performance and notably enhancing safety and security compared to single-objective methods. This advancement marks a significant stride towards the deployment of LLM systems that are both high-performing and secure across varied industrial applications |
format | Article |
fullrecord | <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_3116752519</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3116752519</sourcerecordid><originalsourceid>FETCH-proquest_journals_31167525193</originalsourceid><addsrcrecordid>eNqNzc0KgkAYheEhCIrqHj5oLehMarWNohZRoHuZ8jNHRsfmL-jqM-gCWp3F-8AZkSllLArWK0onZGFME4YhTVIax2xK6sxpLzyXoCqwNULGKzR2C7l6cV0ayPDuNMJVq7a3cOmtaMWbW6G6gWvlHjWcOotaIvdYwtlJK4LLrcG7FR5h75V0Xz0n44pLg4vfzsjysM93x6DX6umGy6JRTndDKlgUJWlM42jD_lMfpU5Ijw</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3116752519</pqid></control><display><type>article</type><title>Survival of the Safest: Towards Secure Prompt Optimization through Interleaved Multi-Objective Evolution</title><source>Free E- Journals</source><creator>Sinha, Ankita ; Cui, Wendi ; Das, Kamalika ; Zhang, Jiaxin</creator><creatorcontrib>Sinha, Ankita ; Cui, Wendi ; Das, Kamalika ; Zhang, Jiaxin</creatorcontrib><description>Large language models (LLMs) have demonstrated remarkable capabilities; however, the optimization of their prompts has historically prioritized performance metrics at the expense of crucial safety and security considerations. To overcome this shortcoming, we introduce "Survival of the Safest" (SoS), an innovative multi-objective prompt optimization framework that enhances both performance and security in LLMs simultaneously. SoS utilizes an interleaved multi-objective evolution strategy, integrating semantic, feedback, and crossover mutations to effectively traverse the prompt landscape. Differing from the computationally demanding Pareto front methods, SoS provides a scalable solution that expedites optimization in complex, high-dimensional discrete search spaces while keeping computational demands low. Our approach accommodates flexible weighting of objectives and generates a pool of optimized candidates, empowering users to select prompts that optimally meet their specific performance and security needs. Experimental evaluations across diverse benchmark datasets affirm SoS's efficacy in delivering high performance and notably enhancing safety and security compared to single-objective methods. This advancement marks a significant stride towards the deployment of LLM systems that are both high-performing and secure across varied industrial applications</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Industrial applications ; Large language models ; Multiple objective analysis ; Optimization ; Performance evaluation ; Performance measurement ; Prompt engineering ; Security ; Survival</subject><ispartof>arXiv.org, 2024-10</ispartof><rights>2024. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>780,784</link.rule.ids></links><search><creatorcontrib>Sinha, Ankita</creatorcontrib><creatorcontrib>Cui, Wendi</creatorcontrib><creatorcontrib>Das, Kamalika</creatorcontrib><creatorcontrib>Zhang, Jiaxin</creatorcontrib><title>Survival of the Safest: Towards Secure Prompt Optimization through Interleaved Multi-Objective Evolution</title><title>arXiv.org</title><description>Large language models (LLMs) have demonstrated remarkable capabilities; however, the optimization of their prompts has historically prioritized performance metrics at the expense of crucial safety and security considerations. To overcome this shortcoming, we introduce "Survival of the Safest" (SoS), an innovative multi-objective prompt optimization framework that enhances both performance and security in LLMs simultaneously. SoS utilizes an interleaved multi-objective evolution strategy, integrating semantic, feedback, and crossover mutations to effectively traverse the prompt landscape. Differing from the computationally demanding Pareto front methods, SoS provides a scalable solution that expedites optimization in complex, high-dimensional discrete search spaces while keeping computational demands low. Our approach accommodates flexible weighting of objectives and generates a pool of optimized candidates, empowering users to select prompts that optimally meet their specific performance and security needs. Experimental evaluations across diverse benchmark datasets affirm SoS's efficacy in delivering high performance and notably enhancing safety and security compared to single-objective methods. This advancement marks a significant stride towards the deployment of LLM systems that are both high-performing and secure across varied industrial applications</description><subject>Industrial applications</subject><subject>Large language models</subject><subject>Multiple objective analysis</subject><subject>Optimization</subject><subject>Performance evaluation</subject><subject>Performance measurement</subject><subject>Prompt engineering</subject><subject>Security</subject><subject>Survival</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNqNzc0KgkAYheEhCIrqHj5oLehMarWNohZRoHuZ8jNHRsfmL-jqM-gCWp3F-8AZkSllLArWK0onZGFME4YhTVIax2xK6sxpLzyXoCqwNULGKzR2C7l6cV0ayPDuNMJVq7a3cOmtaMWbW6G6gWvlHjWcOotaIvdYwtlJK4LLrcG7FR5h75V0Xz0n44pLg4vfzsjysM93x6DX6umGy6JRTndDKlgUJWlM42jD_lMfpU5Ijw</recordid><startdate>20241012</startdate><enddate>20241012</enddate><creator>Sinha, Ankita</creator><creator>Cui, Wendi</creator><creator>Das, Kamalika</creator><creator>Zhang, Jiaxin</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20241012</creationdate><title>Survival of the Safest: Towards Secure Prompt Optimization through Interleaved Multi-Objective Evolution</title><author>Sinha, Ankita ; Cui, Wendi ; Das, Kamalika ; Zhang, Jiaxin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_31167525193</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Industrial applications</topic><topic>Large language models</topic><topic>Multiple objective analysis</topic><topic>Optimization</topic><topic>Performance evaluation</topic><topic>Performance measurement</topic><topic>Prompt engineering</topic><topic>Security</topic><topic>Survival</topic><toplevel>online_resources</toplevel><creatorcontrib>Sinha, Ankita</creatorcontrib><creatorcontrib>Cui, Wendi</creatorcontrib><creatorcontrib>Das, Kamalika</creatorcontrib><creatorcontrib>Zhang, Jiaxin</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Sinha, Ankita</au><au>Cui, Wendi</au><au>Das, Kamalika</au><au>Zhang, Jiaxin</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Survival of the Safest: Towards Secure Prompt Optimization through Interleaved Multi-Objective Evolution</atitle><jtitle>arXiv.org</jtitle><date>2024-10-12</date><risdate>2024</risdate><eissn>2331-8422</eissn><abstract>Large language models (LLMs) have demonstrated remarkable capabilities; however, the optimization of their prompts has historically prioritized performance metrics at the expense of crucial safety and security considerations. To overcome this shortcoming, we introduce "Survival of the Safest" (SoS), an innovative multi-objective prompt optimization framework that enhances both performance and security in LLMs simultaneously. SoS utilizes an interleaved multi-objective evolution strategy, integrating semantic, feedback, and crossover mutations to effectively traverse the prompt landscape. Differing from the computationally demanding Pareto front methods, SoS provides a scalable solution that expedites optimization in complex, high-dimensional discrete search spaces while keeping computational demands low. Our approach accommodates flexible weighting of objectives and generates a pool of optimized candidates, empowering users to select prompts that optimally meet their specific performance and security needs. Experimental evaluations across diverse benchmark datasets affirm SoS's efficacy in delivering high performance and notably enhancing safety and security compared to single-objective methods. This advancement marks a significant stride towards the deployment of LLM systems that are both high-performing and secure across varied industrial applications</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | EISSN: 2331-8422 |
ispartof | arXiv.org, 2024-10 |
issn | 2331-8422 |
language | eng |
recordid | cdi_proquest_journals_3116752519 |
source | Free E- Journals |
subjects | Industrial applications Large language models Multiple objective analysis Optimization Performance evaluation Performance measurement Prompt engineering Security Survival |
title | Survival of the Safest: Towards Secure Prompt Optimization through Interleaved Multi-Objective Evolution |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-08T18%3A51%3A50IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Survival%20of%20the%20Safest:%20Towards%20Secure%20Prompt%20Optimization%20through%20Interleaved%20Multi-Objective%20Evolution&rft.jtitle=arXiv.org&rft.au=Sinha,%20Ankita&rft.date=2024-10-12&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E3116752519%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3116752519&rft_id=info:pmid/&rfr_iscdi=true |