{"id":"https://openalex.org/W4392971579","doi":"https://doi.org/10.48550/arxiv.2403.11838","title":"Ensuring Safe and High-Quality Outputs: A Guideline Library Approach for Language Models","display_name":"Ensuring Safe and High-Quality Outputs: A Guideline Library Approach for Language Models","publication_year":2024,"publication_date":"2024-03-18","ids":{"openalex":"https://openalex.org/W4392971579","doi":"https://doi.org/10.48550/arxiv.2403.11838"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2403.11838","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2403.11838","pdf_url":"https://arxiv.org/pdf/2403.11838","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2403.11838","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5054109175","display_name":"Yi Luo","orcid":"https://orcid.org/0000-0002-6877-9327"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Luo, Yi","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101243010","display_name":"Zhenghao Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Zhenghao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100349352","display_name":"Yuhao Zhang","orcid":"https://orcid.org/0000-0002-9856-436X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yuhao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102632633","display_name":"Jiashuo Sun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Jiashuo","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5082333799","display_name":"Chen Lin","orcid":"https://orcid.org/0000-0003-2471-5222"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Chen","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022319543","display_name":"Chengjin Xu","orcid":"https://orcid.org/0000-0002-4942-9049"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Chengjin","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010464164","display_name":"Xiangdong Su","orcid":"https://orcid.org/0000-0001-8061-1474"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Su, Xiangdong","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101180037","display_name":"Yelong Shen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Yelong","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100756159","display_name":"Jian Guo","orcid":"https://orcid.org/0000-0002-4394-5560"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Jian","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5041448669","display_name":"Yeyun Gong","orcid":"https://orcid.org/0000-0001-9954-9674"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gong, Yeyun","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":10,"corresponding_author_ids":["https://openalex.org/A5054109175"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.6320000290870667,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.6320000290870667,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.6272000074386597,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/guideline","display_name":"Guideline","score":0.7209365367889404},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.603843092918396},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5539080500602722},{"id":"https://openalex.org/keywords/risk-analysis","display_name":"Risk analysis (engineering)","score":0.40192872285842896},{"id":"https://openalex.org/keywords/process-management","display_name":"Process management","score":0.3603070378303528},{"id":"https://openalex.org/keywords/business","display_name":"Business","score":0.32589733600616455},{"id":"https://openalex.org/keywords/political-science","display_name":"Political science","score":0.17106389999389648}],"concepts":[{"id":"https://openalex.org/C2780182762","wikidata":"https://www.wikidata.org/wiki/Q1630279","display_name":"Guideline","level":2,"score":0.7209365367889404},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.603843092918396},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5539080500602722},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.40192872285842896},{"id":"https://openalex.org/C195094911","wikidata":"https://www.wikidata.org/wiki/Q14167904","display_name":"Process management","level":1,"score":0.3603070378303528},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.32589733600616455},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.17106389999389648},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.0},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2403.11838","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2403.11838","pdf_url":"https://arxiv.org/pdf/2403.11838","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2403.11838","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2403.11838","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2403.11838","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2403.11838","pdf_url":"https://arxiv.org/pdf/2403.11838","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G2087396116","display_name":null,"funder_award_id":"China","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G3317480652","display_name":null,"funder_award_id":"Science","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G56815992","display_name":null,"funder_award_id":"62372390","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G5994120800","display_name":null,"funder_award_id":"Natural","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4392971579.pdf"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2339422290","https://openalex.org/W2397653702","https://openalex.org/W4233871529","https://openalex.org/W2338818750","https://openalex.org/W2365451219","https://openalex.org/W2415869793","https://openalex.org/W4211072166","https://openalex.org/W28723353","https://openalex.org/W4231104876","https://openalex.org/W2441306059"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"exhibit":[4],"impressive":[5],"capabilities":[6],"but":[7,28],"also":[8],"present":[9],"risks":[10,64],"such":[11],"as":[12],"biased":[13],"content":[14],"generation":[15,99],"and":[16,40,65,78,103,143,176,190],"privacy":[17],"issues.":[18],"One":[19],"of":[20,36,76,145],"the":[21,34,85,124,128,141,146],"current":[22],"alignment":[23,194],"techniques":[24],"includes":[25],"principle-driven":[26],"integration,":[27],"it":[29,150],"faces":[30],"challenges":[31],"arising":[32],"from":[33,154],"imprecision":[35],"manually":[37],"crafted":[38],"rules":[39],"inadequate":[41],"risk":[42],"perception":[43],"in":[44,97,127,173,193],"models":[45],"without":[46],"safety":[47,152],"training.":[48],"To":[49],"address":[50],"these,":[51],"we":[52],"introduce":[53],"Guide-Align,":[54],"a":[55,59,73,79,117,155,159],"two-stage":[56],"approach.":[57],"Initially,":[58],"safety-trained":[60,156],"model":[61,80,87,118],"identifies":[62],"potential":[63],"formulates":[66],"specific":[67],"guidelines":[68,77,134],"for":[69,81],"various":[70],"inputs,":[71,138],"establishing":[72],"comprehensive":[74],"library":[75],"input-guidelines":[82],"retrieval.":[83],"Subsequently,":[84],"retrieval":[86,161],"correlates":[88],"new":[89],"inputs":[90],"with":[91,108,119],"relevant":[92],"guidelines,":[93],"which":[94],"guide":[95],"LLMs":[96],"response":[98],"to":[100,135],"ensure":[101],"safe":[102],"high-quality":[104],"outputs,":[105],"thereby":[106,139],"aligning":[107],"human":[109],"values.":[110],"An":[111],"additional":[112],"optional":[113],"stage":[114],"involves":[115],"fine-tuning":[116],"well-aligned":[120],"datasets":[121],"generated":[122],"through":[123,158],"process":[125],"implemented":[126],"second":[129],"stage.":[130],"Our":[131],"method":[132],"customizes":[133],"accommodate":[136],"diverse":[137],"enhancing":[140],"fine-grainedness":[142],"comprehensiveness":[144],"guideline":[147],"library.":[148],"Furthermore,":[149],"incorporates":[151],"expertise":[153],"LLM":[157,174],"lightweight":[160],"model.":[162],"We":[163],"evaluate":[164],"our":[165,179],"approach":[166],"on":[167],"three":[168],"benchmarks,":[169],"demonstrating":[170],"significant":[171],"improvements":[172],"security":[175],"quality.":[177],"Notably,":[178],"fine-tuned":[180],"model,":[181],"Labrador,":[182],"even":[183],"at":[184],"13":[185],"billion":[186],"parameters,":[187],"outperforms":[188],"GPT-3.5-turbo":[189],"surpasses":[191],"GPT-4":[192],"capabilities.":[195]},"counts_by_year":[],"updated_date":"2026-04-18T07:56:08.524223","created_date":"2025-10-10T00:00:00"}