{"id":"https://openalex.org/W7106767948","doi":"https://doi.org/10.48550/arxiv.2511.19399","title":"DR Tulu: Reinforcement Learning with Evolving Rubrics for Deep Research","display_name":"DR Tulu: Reinforcement Learning with Evolving Rubrics for Deep Research","publication_year":2025,"publication_date":"2025-11-24","ids":{"openalex":"https://openalex.org/W7106767948","doi":"https://doi.org/10.48550/arxiv.2511.19399"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2511.19399","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2511.19399","pdf_url":"https://arxiv.org/pdf/2511.19399","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","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2511.19399","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Shao, Rulin","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shao, Rulin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Asai, Akari","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Asai, Akari","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Shen, Shannon Zejiang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Shannon Zejiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Ivison, Hamish","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ivison, Hamish","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Kishore, Varsha","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kishore, Varsha","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zhuo, Jingming","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhuo, Jingming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102892146","display_name":"Xinran Zhao","orcid":"https://orcid.org/0009-0006-0938-7602","observed_orcids":["https://orcid.org/0009-0006-0938-7602"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Xinran","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Park, Molly","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Park, Molly","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090442121","display_name":"Samuel G. Finlayson","orcid":"https://orcid.org/0000-0001-7997-5959","observed_orcids":["https://orcid.org/0000-0001-7997-5959"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Finlayson, Samuel G.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013431623","display_name":"David Alexander Sontag","orcid":"https://orcid.org/0000-0002-5034-7796","observed_orcids":["https://orcid.org/0000-0002-5034-7796","https://orcid.org/0000-0002-7798-2925"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sontag, David","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Murray, Tyler","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Murray, Tyler","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Min, Sewon","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Min, Sewon","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Dasigi, Pradeep","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dasigi, Pradeep","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Soldaini, Luca","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Soldaini, Luca","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Brahman, Faeze","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Brahman, Faeze","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yih, Wen-tau","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yih, Wen-tau","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004225142","display_name":"Tongshuang Wu","orcid":"https://orcid.org/0000-0003-1630-0588","observed_orcids":["https://orcid.org/0000-0003-1630-0588"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Tongshuang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067919401","display_name":"Luke Zettlemoyer","orcid":"https://orcid.org/0009-0008-8296-0764","observed_orcids":["https://orcid.org/0009-0008-8296-0764"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zettlemoyer, Luke","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Kim, Yoon","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kim, Yoon","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Hajishirzi, Hannaneh","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hajishirzi, Hannaneh","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5079904764","display_name":"Pang Wei Koh","orcid":"https://orcid.org/0000-0003-4330-6969","observed_orcids":["https://orcid.org/0000-0003-4330-6969"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Koh, Pang Wei","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":false,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.42089998722076416,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.42089998722076416,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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":0.35690000653266907,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.039900001138448715,"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/deep-research-agents","display_name":"deep research agents","score":0.9330000281333923},{"id":"https://openalex.org/keywords/fact-checking","display_name":"fact-checking","score":0.8880000114440918},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.6859999895095825}],"concepts":[{"id":"https://openalex.org/C111640148","wikidata":"https://www.wikidata.org/wiki/Q847349","display_name":"Rubric","level":2,"score":0.921999990940094},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.7522000074386597},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.7008000016212463},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6291000247001648},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.611299991607666},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5949000120162964},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.5217000246047974},{"id":"https://openalex.org/C2778464652","wikidata":"https://www.wikidata.org/wiki/Q309849","display_name":"Open research","level":2,"score":0.41290000081062317},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.36149999499320984},{"id":"https://openalex.org/C2777826224","wikidata":"https://www.wikidata.org/wiki/Q5250404","display_name":"Deep integration","level":2,"score":0.2955999970436096},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.29010000824928284},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.2815000116825104}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2511.19399","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2511.19399","pdf_url":"https://arxiv.org/pdf/2511.19399","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.2511.19399","doi":"https://doi.org/10.48550/arxiv.2511.19399","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2511.19399","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:2511.19399","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2511.19399","pdf_url":"https://arxiv.org/pdf/2511.19399","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":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7160937593","https://openalex.org/W4416189150","https://openalex.org/W7164941083","https://openalex.org/W7163269026","https://openalex.org/W7131424306","https://openalex.org/W4415845669","https://openalex.org/W4416037139","https://openalex.org/W7102388436","https://openalex.org/W7123371307","https://openalex.org/W7125790422"],"abstract_inverted_index":{"Deep":[0,91],"research":[1,5,15,114,129,144],"agents":[2,16,130,145],"perform":[3],"multi-step":[4],"to":[6,35,55,67],"produce":[7],"long-form,":[8],"well-attributed":[9],"answers.":[10],"However,":[11],"most":[12],"open":[13,99,127],"deep":[14,108,113,128,143],"are":[17,51],"trained":[18,104],"on":[19,136,151],"easily":[20],"verifiable":[21,29],"short-form":[22],"QA":[23],"tasks":[24],"via":[25],"reinforcement":[26],"learning":[27],"with":[28,42,45,57],"rewards,":[30],"which":[31],"does":[32],"not":[33],"extend":[34],"realistic":[36],"long-form":[37,107,112],"tasks.":[38],"We":[39],"address":[40],"this":[41],"Reinforcement":[43],"Learning":[44],"Evolving":[46],"Rubrics":[47],"(RLER),":[48],"where":[49],"rubrics":[50,66],"constructed":[52],"and":[53,74,82,119,138,157],"maintained":[54],"co-evolve":[56],"the":[58,65,96],"policy":[59],"model":[60,76,100],"during":[61],"training.":[62],"This":[63],"allows":[64],"incorporate":[68],"newly":[69],"explored":[70],"information":[71],"from":[72],"search":[73],"contrasting":[75],"responses,":[77],"enabling":[78],"better":[79],"fact":[80],"checking":[81],"more":[83],"discriminative":[84],"on-policy":[85],"feedback.":[86],"Using":[87],"RLER,":[88],"we":[89],"develop":[90],"Research":[92],"Tulu":[93,123],"(DR":[94],"Tulu-8B),":[95],"first":[97],"fully":[98],"that":[101],"is":[102],"directly":[103],"for":[105],"open-ended,":[106],"research.":[109],"Across":[110],"four":[111],"benchmarks":[115],"in":[116],"science,":[117],"healthcare,":[118],"general":[120],"domains,":[121],"DR":[122,135,150,165],"substantially":[124],"outperforms":[125],"existing":[126],"(by":[131,146],"15.6%":[132],"over":[133,148],"Tongyi":[134],"average)":[137],"matches":[139],"or":[140],"exceeds":[141],"proprietary":[142],"0.7%":[147],"OpenAI":[149,164],"average),":[152],"while":[153],"being":[154],"significantly":[155],"smaller":[156],"cheaper":[158,162],"per":[159,166],"query":[160],"(1000x":[161],"than":[163],"query).":[167]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2025-11-27T00:00:00"}