{"id":"https://openalex.org/W7154171807","doi":"https://doi.org/10.48550/arxiv.2604.09494","title":"RecaLLM: Addressing the Lost-in-Thought Phenomenon with Explicit In-Context Retrieval","display_name":"RecaLLM: Addressing the Lost-in-Thought Phenomenon with Explicit In-Context Retrieval","publication_year":2026,"publication_date":"2026-04-10","ids":{"openalex":"https://openalex.org/W7154171807","doi":"https://doi.org/10.48550/arxiv.2604.09494"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.09494","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.09494","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,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.09494","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5091285923","display_name":"Kyle Whitecross","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Whitecross, Kyle","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5118214885","display_name":"Negin Rahimi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rahimi, Negin","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.5444999933242798,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.5444999933242798,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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":0.27149999141693115,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.0494999997317791,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.559499979019165},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5562000274658203},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.5358999967575073},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.40119999647140503},{"id":"https://openalex.org/keywords/encoding","display_name":"Encoding (memory)","score":0.37869998812675476},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.37439998984336853},{"id":"https://openalex.org/keywords/question-answering","display_name":"Question answering","score":0.374099999666214},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.35929998755455017},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.3474000096321106}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8310999870300293},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5651000142097473},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.559499979019165},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5562000274658203},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.5358999967575073},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4300000071525574},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.40119999647140503},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.37869998812675476},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.37439998984336853},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.374099999666214},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.35929998755455017},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.35420000553131104},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.3474000096321106},{"id":"https://openalex.org/C2779151265","wikidata":"https://www.wikidata.org/wiki/Q1156791","display_name":"Copying","level":2,"score":0.34459999203681946},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.3416999876499176},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.33009999990463257},{"id":"https://openalex.org/C100660578","wikidata":"https://www.wikidata.org/wiki/Q18733","display_name":"Recall","level":2,"score":0.32580000162124634},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.2957000136375427},{"id":"https://openalex.org/C89288958","wikidata":"https://www.wikidata.org/wiki/Q7301504","display_name":"Reasoning system","level":2,"score":0.2775999903678894},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2727999985218048},{"id":"https://openalex.org/C20162079","wikidata":"https://www.wikidata.org/wiki/Q1151406","display_name":"Case-based reasoning","level":2,"score":0.2727000117301941},{"id":"https://openalex.org/C193221554","wikidata":"https://www.wikidata.org/wiki/Q5153664","display_name":"Commonsense reasoning","level":2,"score":0.2685999870300293},{"id":"https://openalex.org/C50335755","wikidata":"https://www.wikidata.org/wiki/Q483247","display_name":"Phenomenon","level":2,"score":0.2648000121116638},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.2632000148296356},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.2630999982357025},{"id":"https://openalex.org/C83725634","wikidata":"https://www.wikidata.org/wiki/Q7268699","display_name":"Qualitative reasoning","level":2,"score":0.2587999999523163},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.2542000114917755}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.09494","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.09494","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.09494","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.09494","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,"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"We":[0,116],"propose":[1],"RecaLLM,":[2],"a":[3,64,69,118,189],"set":[4],"of":[5,14,127,133,166,174],"reasoning":[6,26,34,66,81,99,106],"language":[7],"models":[8],"post-trained":[9],"to":[10,78,112,168],"make":[11,87],"effective":[12],"use":[13],"long-context":[15,150,186,194,198],"information.":[16],"In-context":[17],"retrieval,":[18,103],"which":[19],"identifies":[20],"relevant":[21],"evidence":[22,128],"from":[23],"context,":[24],"and":[25,107,140,153],"are":[27],"deeply":[28],"intertwined:":[29],"retrieval":[30,58,90,142],"supports":[31],"reasoning,":[32],"while":[33],"often":[35],"determines":[36],"what":[37],"must":[38],"be":[39],"retrieved.":[40],"However,":[41],"their":[42],"interaction":[43],"remains":[44],"largely":[45],"underexplored.":[46],"In":[47],"preliminary":[48],"experiments":[49],"on":[50,137,148],"several":[51],"open-source":[52],"LLMs,":[53],"we":[54,76,159],"observe":[55,160],"that":[56,75,83,123],"in-context":[57,89,102],"performance":[59,85,147,195],"substantially":[60],"degrades":[61],"even":[62],"after":[63],"short":[65],"span,":[67],"revealing":[68],"key":[70],"bottleneck":[71],"for":[72],"test-time":[73],"scaling":[74],"refer":[77],"as":[79],"lost-in-thought:":[80],"steps":[82],"improve":[84],"also":[86],"subsequent":[88,134],"more":[91],"challenging.":[92],"To":[93],"address":[94],"this":[95],"limitation,":[96],"RecaLLM":[97,144],"interleaves":[98],"with":[100],"explicit":[101],"alternating":[104],"between":[105],"retrieving":[108],"context":[109,164],"information":[110],"needed":[111],"solve":[113],"intermediate":[114],"subproblems.":[115],"introduce":[117],"negligible-overhead":[119],"constrained":[120],"decoding":[121],"mechanism":[122],"enables":[124],"verbatim":[125],"copying":[126],"spans,":[129],"improving":[130,193],"the":[131],"grounding":[132],"generation.":[135],"Trained":[136],"diverse":[138],"lexical":[139],"semantic":[141],"tasks,":[143],"achieves":[145],"strong":[146],"two":[149],"benchmarks,":[151],"RULER":[152],"HELMET,":[154],"significantly":[155],"outperforming":[156],"baselines.":[157],"Notably,":[158],"consistent":[161],"gains":[162],"at":[163,175],"windows":[165],"up":[167],"128K":[169],"tokens":[170],"using":[171],"training":[172,199],"samples":[173],"most":[176],"10K":[177],"tokens,":[178],"far":[179],"shorter":[180],"than":[181],"those":[182],"used":[183],"by":[184],"existing":[185],"approaches,":[187],"highlighting":[188],"promising":[190],"path":[191],"toward":[192],"without":[196],"expensive":[197],"data.":[200]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-04-14T00:00:00"}