{"id":"https://openalex.org/W4403573990","doi":"https://doi.org/10.48550/arxiv.2410.11348","title":"RATE: Causal Explainability of Reward Models with Imperfect Counterfactuals","display_name":"RATE: Causal Explainability of Reward Models with Imperfect Counterfactuals","publication_year":2024,"publication_date":"2024-10-15","ids":{"openalex":"https://openalex.org/W4403573990","doi":"https://doi.org/10.48550/arxiv.2410.11348"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2410.11348","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2410.11348","pdf_url":"https://arxiv.org/pdf/2410.11348","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2410.11348","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Reber, David","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Reber, David","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104594024","display_name":"Sean Richardson","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Richardson, Sean","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114336283","display_name":"Todd Nief","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nief, Todd","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028298718","display_name":"Cristina G\u00e2rbacea","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Garbacea, Cristina","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5077406767","display_name":"Victor Veitch","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Veitch, Victor","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":6,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9628999829292297,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9628999829292297,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.014700000174343586,"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":0.0027000000700354576,"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/reward-models","display_name":"reward models","score":0.9829999804496765},{"id":"https://openalex.org/keywords/counterfactuals","display_name":"counterfactuals","score":0.6200000047683716},{"id":"https://openalex.org/keywords/causal-effect-estimation","display_name":"causal effect estimation","score":0.6480000019073486},{"id":"https://openalex.org/keywords/bias-correction","display_name":"bias correction","score":0.5669999718666077},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.7620000243186951},{"id":"https://openalex.org/keywords/human-preference-alignment","display_name":"human preference alignment","score":0.49000000953674316}],"concepts":[{"id":"https://openalex.org/C2780310539","wikidata":"https://www.wikidata.org/wiki/Q12547192","display_name":"Imperfect","level":2,"score":0.4870775043964386},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4697932004928589},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.39597031474113464},{"id":"https://openalex.org/C70721500","wikidata":"https://www.wikidata.org/wiki/Q177005","display_name":"Computational biology","level":1,"score":0.39456677436828613},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.29765474796295166},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2274557650089264},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","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:2410.11348","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2410.11348","pdf_url":"https://arxiv.org/pdf/2410.11348","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2410.11348","doi":"https://doi.org/10.48550/arxiv.2410.11348","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2410.11348","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2410.11348","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2410.11348","pdf_url":"https://arxiv.org/pdf/2410.11348","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G8876996369","display_name":null,"funder_award_id":"N00014","funder_id":"https://openalex.org/F4320337345","funder_display_name":"Office of Naval Research"}],"funders":[{"id":"https://openalex.org/F4320315389","display_name":"Open Philanthropy Project","ror":"https://ror.org/004d1k391"},{"id":"https://openalex.org/F4320337345","display_name":"Office of Naval Research","ror":"https://ror.org/00rk2pe57"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4403573990.pdf","grobid_xml":"https://content.openalex.org/works/W4403573990.grobid-xml"},"referenced_works_count":10,"referenced_works":["https://openalex.org/W2950681488","https://openalex.org/W4224005198","https://openalex.org/W4302010373","https://openalex.org/W4312516176","https://openalex.org/W4366196653","https://openalex.org/W4385573536","https://openalex.org/W4385574060","https://openalex.org/W4387323731","https://openalex.org/W4393063868","https://openalex.org/W4406457873"],"related_works":["https://openalex.org/W4417538704","https://openalex.org/W7166807385","https://openalex.org/W7214736017","https://openalex.org/W7158944862","https://openalex.org/W7214809027","https://openalex.org/W4406549323","https://openalex.org/W4407809010","https://openalex.org/W4417097574","https://openalex.org/W7130342037","https://openalex.org/W7139936135"],"abstract_inverted_index":{"Reward":[0],"models":[1,17],"are":[2,18,29,103],"widely":[3],"used":[4,91],"as":[5,42,60],"proxies":[6],"for":[7,46,132],"human":[8],"preferences":[9],"when":[10],"aligning":[11],"or":[12,63],"evaluating":[13],"LLMs.":[14],"However,":[15],"reward":[16,52,119],"black":[19],"boxes,":[20],"and":[21,147],"it":[22,151],"is":[23,99,129,152],"often":[24],"unclear":[25],"what,":[26],"exactly,":[27],"they":[28],"actually":[30],"rewarding.":[31],"In":[32],"this":[33,133],"paper":[34],"we":[35],"develop":[36],"Rewrite-based":[37],"Attribute":[38],"Treatment":[39],"Estimator":[40],"(RATE)":[41],"an":[43,72,153],"effective":[44,154],"method":[45],"measuring":[47],"the":[48,68,75,114,118,122,141,144],"sensitivity":[49,116],"of":[50,57,71,117,127,143],"a":[51,106],"model":[53,120],"to":[54,80,83,92,121,130],"high-level":[55],"attributes":[56],"responses,":[58],"such":[59],"sentiment,":[61],"helpfulness,":[62],"complexity.":[64],"Importantly,":[65],"RATE":[66,77,128,145],"measures":[67],"causal":[69,94],"effect":[70,135],"attribute":[73],"on":[74],"reward.":[76],"uses":[78],"LLMs":[79],"rewrite":[81],"responses":[82],"produce":[84],"imperfect":[85,104],"counterfactuals":[86],"examples":[87],"that":[88,100,108,150],"can":[89,109],"be":[90],"measure":[93],"effects.":[95],"A":[96],"key":[97],"challenge":[98],"these":[101],"rewrites":[102],"in":[105,113],"manner":[107],"induce":[110],"substantial":[111],"bias":[112],"estimated":[115],"attribute.":[123],"The":[124],"core":[125],"idea":[126],"adjust":[131],"imperfect-rewrite":[134],"by":[136],"rewriting":[137],"twice.":[138],"We":[139],"establish":[140],"validity":[142],"procedure":[146],"show":[148],"empirically":[149],"estimator.":[155]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":4}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2024-10-20T00:00:00"}