{"id":"https://openalex.org/W4414744284","doi":"https://doi.org/10.48550/arxiv.2509.16422","title":"Evaluating CxG Generalisation in LLMs via Construction-Based NLI Fine Tuning","display_name":"Evaluating CxG Generalisation in LLMs via Construction-Based NLI Fine Tuning","publication_year":2025,"publication_date":"2025-09-19","ids":{"openalex":"https://openalex.org/W4414744284","doi":"https://doi.org/10.48550/arxiv.2509.16422"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2509.16422","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2509.16422","pdf_url":"https://arxiv.org/pdf/2509.16422","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":null,"license_id":null,"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/2509.16422","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5119812170","display_name":"Tom Mackintosh","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mackintosh, Tom","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070941491","display_name":"Harish Tayyar Madabushi","orcid":"https://orcid.org/0000-0001-5260-3653","observed_orcids":["https://orcid.org/0000-0001-5260-3653"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Madabushi, Harish Tayyar","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5087092493","display_name":"Claire Bonial","orcid":"https://orcid.org/0000-0002-3154-2852","observed_orcids":["https://orcid.org/0000-0002-3154-2852"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bonial, Claire","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/T10028","display_name":"Topic Modeling","score":0.41530001163482666,"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/T10028","display_name":"Topic Modeling","score":0.41530001163482666,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.2842000126838684,"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/T10034","display_name":"Syntax, Semantics, Linguistic Variation","score":0.23680000007152557,"subfield":{"id":"https://openalex.org/subfields/1203","display_name":"Language and Linguistics"},"field":{"id":"https://openalex.org/fields/12","display_name":"Arts and Humanities"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/construction-grammar","display_name":"construction grammar","score":0.9340000152587891},{"id":"https://openalex.org/keywords/natural-language-inference","display_name":"natural language inference","score":0.7699999809265137},{"id":"https://openalex.org/keywords/adversarial-data","display_name":"adversarial data","score":0.5099999904632568},{"id":"https://openalex.org/keywords/llm-fine-tuning","display_name":"LLM fine-tuning","score":0.7699999809265137},{"id":"https://openalex.org/keywords/form-meaning-mapping","display_name":"form-meaning mapping","score":0.5690000057220459}],"concepts":[{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.6553000211715698},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6025999784469604},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5927000045776367},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.5648000240325928},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5130000114440918},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.4970000088214874},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4440999925136566},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.39750000834465027},{"id":"https://openalex.org/C157524613","wikidata":"https://www.wikidata.org/wiki/Q2828883","display_name":"Fine-tuning","level":2,"score":0.38999998569488525},{"id":"https://openalex.org/C192328126","wikidata":"https://www.wikidata.org/wiki/Q4514647","display_name":"Schematic","level":2,"score":0.3776000142097473},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.33739998936653137},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.31610000133514404},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.30889999866485596},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.29409998655319214},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.28529998660087585},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2840000092983246},{"id":"https://openalex.org/C31170391","wikidata":"https://www.wikidata.org/wiki/Q188619","display_name":"Hierarchy","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C97256817","wikidata":"https://www.wikidata.org/wiki/Q1462316","display_name":"Spurious relationship","level":2,"score":0.2603999972343445},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.2572000026702881}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2509.16422","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2509.16422","pdf_url":"https://arxiv.org/pdf/2509.16422","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2509.16422","doi":"https://doi.org/10.48550/arxiv.2509.16422","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2509.16422","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:2509.16422","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2509.16422","pdf_url":"https://arxiv.org/pdf/2509.16422","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":null,"license_id":null,"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":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4406231702","https://openalex.org/W4405617014","https://openalex.org/W7131796284","https://openalex.org/W7077161912","https://openalex.org/W7160087855","https://openalex.org/W7166855009","https://openalex.org/W7160241332","https://openalex.org/W7111984276","https://openalex.org/W4415593199","https://openalex.org/W4414786602"],"abstract_inverted_index":{"We":[0,16],"probe":[1],"large":[2],"language":[3],"models'":[4],"ability":[5],"to":[6,31,56,93],"learn":[7],"deep":[8],"form-meaning":[9],"mappings":[10],"as":[11],"defined":[12],"by":[13],"construction":[14],"grammars.":[15],"introduce":[17],"the":[18,44],"ConTest-NLI":[19,90],"benchmark":[20],"of":[21,46,53,89],"80k":[22],"sentences":[23],"covering":[24],"eight":[25],"English":[26],"constructions":[27],"from":[28],"highly":[29,32],"lexicalized":[30],"schematic.":[33],"Our":[34],"pipeline":[35],"generates":[36],"diverse":[37],"synthetic":[38],"NLI":[39],"triples":[40],"via":[41],"templating":[42],"and":[43,59,76,106],"application":[45],"a":[47,68,87,108],"model-in-the-loop":[48],"filter.":[49],"This":[50],"provides":[51],"aspects":[52],"human":[54],"validation":[55],"ensure":[57],"challenge":[58],"label":[60],"reliability.":[61],"Zero-shot":[62],"tests":[63],"on":[64,86],"leading":[65],"LLMs":[66,105],"reveal":[67],"24%":[69],"drop":[70],"in":[71,103],"accuracy":[72],"between":[73],"naturalistic":[74],"(88%)":[75],"adversarial":[77],"data":[78],"(64%),":[79],"with":[80],"schematic":[81],"patterns":[82],"proving":[83],"hardest.":[84],"Fine-tuning":[85],"subset":[88],"yields":[91],"up":[92],"9%":[94],"improvement,":[95],"yet":[96],"our":[97],"results":[98],"highlight":[99],"persistent":[100],"abstraction":[101],"gaps":[102],"current":[104],"offer":[107],"scalable":[109],"framework":[110],"for":[111],"evaluating":[112],"construction-informed":[113],"learning.":[114]},"counts_by_year":[],"updated_date":"2026-08-26T07:47:46.906454","created_date":"2025-10-10T00:00:00"}