{"id":"https://openalex.org/W4416037520","doi":"https://doi.org/10.18653/v1/2025.emnlp-demos.62","title":"TRACE: Training and Inference-Time Interpretability Analysis for Language Models","display_name":"TRACE: Training and Inference-Time Interpretability Analysis for Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416037520","doi":"https://doi.org/10.18653/v1/2025.emnlp-demos.62"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.emnlp-demos.62","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-demos.62","pdf_url":"https://aclanthology.org/2025.emnlp-demos.62.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.emnlp-demos.62.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5114337749","display_name":"Nura Aljaafari","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nura Aljaafari","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Danilo Carvalho","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Danilo Carvalho","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5022138909","display_name":"Andr\u00e9 Victor Lucci Freitas","orcid":"https://orcid.org/0000-0002-5763-4990"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Andre Freitas","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.30918254,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"806","last_page":"820"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.358599990606308,"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":0.358599990606308,"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":0.1290999948978424,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.06599999964237213,"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/interpretability","display_name":"Interpretability","score":0.8614000082015991},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.5110999941825867},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.41359999775886536},{"id":"https://openalex.org/keywords/interpretation","display_name":"Interpretation (philosophy)","score":0.3077000081539154},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.30730000138282776}],"concepts":[{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.8614000082015991},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6169999837875366},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5641999840736389},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.5110999941825867},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5065000057220459},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.41359999775886536},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.32249999046325684},{"id":"https://openalex.org/C527412718","wikidata":"https://www.wikidata.org/wiki/Q855395","display_name":"Interpretation (philosophy)","level":2,"score":0.3077000081539154},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.30730000138282776},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3012999892234802},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.2978000044822693},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.29589998722076416}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.18653/v1/2025.emnlp-demos.62","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-demos.62","pdf_url":"https://aclanthology.org/2025.emnlp-demos.62.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations","raw_type":"proceedings-article"},{"id":"pmh:oai:pure.atira.dk:openaire/2109e59b-3223-4719-8c9d-4dc241acb400","is_oa":true,"landing_page_url":"https://research.manchester.ac.uk/en/publications/2109e59b-3223-4719-8c9d-4dc241acb400","pdf_url":null,"source":{"id":"https://openalex.org/S4306400662","display_name":"Research Explorer (The University of Manchester)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I28407311","host_organization_name":"University of Manchester","host_organization_lineage":["https://openalex.org/I28407311"],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Aljaafari, N, Carvalho, D & Freitas, A 2025, TRACE: Training and Inference-Time Interpretability Analysis for Language Models. in Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations. https://doi.org/10.18653/v1/2025.emnlp-demos.62","raw_type":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.emnlp-demos.62","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-demos.62","pdf_url":"https://aclanthology.org/2025.emnlp-demos.62.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320319985","display_name":"Cancer Research UK","ror":"https://ror.org/054225q67"},{"id":"https://openalex.org/F4320319990","display_name":"National Institute for Health and Care Research","ror":"https://ror.org/0187kwz08"},{"id":"https://openalex.org/F4320320924","display_name":"Schweizerischer Nationalfonds zur F\u00f6rderung der Wissenschaftlichen Forschung","ror":"https://ror.org/00yjd3n13"},{"id":"https://openalex.org/F4320333687","display_name":"Manchester Biomedical Research Centre","ror":null}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416037520.pdf","grobid_xml":"https://content.openalex.org/works/W4416037520.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Understanding":[0],"when":[1],"and":[2,38,47,60,70,109,134,144],"how":[3],"linguistic":[4,59,89],"knowledge":[5],"emerges":[6],"during":[7],"language":[8],"model":[9],"training":[10,46],"remains":[11],"a":[12,42,76],"central":[13],"challenge":[14],"for":[15,45,85],"interpretability.Most":[16],"existing":[17],"tools":[18],"are":[19],"post":[20],"hoc,":[21],"rely":[22],"on":[23],"scalar":[24,116],"metrics,":[25],"or":[26,121],"require":[27],"nontrivial":[28],"integration":[29,124],"effort,":[30,125],"making":[31,139],"comprehensive":[32],"interpretability":[33,49],"analysis":[34,50,57,141],"difficult":[35],"to":[36],"deploy":[37],"maintain.We":[39],"introduce":[40],"TRACE,":[41],"modular":[43],"toolkit":[44],"inference-time":[48],"of":[51,58,88,136],"transformer":[52,140],"models.It":[53],"enables":[54,128],"lightweight,":[55],"in-training":[56],"representational":[61,110],"signals,":[62],"including":[63],"features":[64],"probing,":[65],"intrinsic":[66],"dimensionality,":[67],"Hessian":[68],"curvature,":[69],"output":[71],"diagnostics.It":[72],"integrates":[73],"with":[74,92],"ABSynth,":[75],"controllable":[77],"synthetic":[78],"corpus":[79],"generator":[80],"that":[81,96],"provides":[82],"structured":[83],"annotations":[84],"precise":[86],"evaluation":[87],"feature":[90],"acquisition.Experiments":[91],"autoregressive":[93],"transformers":[94],"demonstrate":[95],"TRACE":[97],"reveals":[98],"developmental":[99],"phenomena":[100],"such":[101,118],"as":[102,119],"early":[103,132],"syntactic":[104],"emergence,":[105],"delayed":[106],"semantic":[107],"acquisition,":[108],"compression,":[111],"signals":[112],"overlooked":[113],"by":[114],"traditional":[115],"metrics":[117],"loss":[120],"accuracy.With":[122],"minimal":[123],"the":[126],"tool":[127],"layer-wise":[129],"diagnostics,":[130],"convergencebased":[131],"stopping,":[133],"detection":[135],"structural":[137],"errors,":[138],"interpretable,":[142],"actionable,":[143],"reproducible.":[145]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-11-08T00:00:00"}