{"id":"https://openalex.org/W4404573447","doi":"https://doi.org/10.48550/arxiv.2411.11984","title":"Understanding Chain-of-Thought in LLMs through Information Theory","display_name":"Understanding Chain-of-Thought in LLMs through Information Theory","publication_year":2024,"publication_date":"2024-11-18","ids":{"openalex":"https://openalex.org/W4404573447","doi":"https://doi.org/10.48550/arxiv.2411.11984"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2411.11984","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.11984","pdf_url":"https://arxiv.org/pdf/2411.11984","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/2411.11984","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5069983126","display_name":"Jean-Fran\u00e7ois Ton","orcid":"https://orcid.org/0000-0002-2756-8945","observed_orcids":["https://orcid.org/0000-0002-2756-8945"]},"institutions":[{"id":"https://openalex.org/I185103710","display_name":"University of California, Santa Cruz","ror":"https://ror.org/03s65by71","country_code":"US","type":"education","lineage":["https://openalex.org/I185103710"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ton, Jean-Francois","raw_affiliation_strings":["University of California , Santa Cruz"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of California , Santa Cruz","institution_ids":["https://openalex.org/I185103710"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Taufiq, Muhammad Faaiz","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I185103710","display_name":"University of California, Santa Cruz","ror":"https://ror.org/03s65by71","country_code":"US","type":"education","lineage":["https://openalex.org/I185103710"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Taufiq, Muhammad Faaiz","raw_affiliation_strings":["University of California , Santa Cruz"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of California , Santa Cruz","institution_ids":["https://openalex.org/I185103710"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101767146","display_name":"Liu Yang","orcid":"https://orcid.org/0000-0002-7230-6485","observed_orcids":["https://orcid.org/0000-0002-7230-6485"]},"institutions":[{"id":"https://openalex.org/I185103710","display_name":"University of California, Santa Cruz","ror":"https://ror.org/03s65by71","country_code":"US","type":"education","lineage":["https://openalex.org/I185103710"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Liu, Yang","raw_affiliation_strings":["University of California , Santa Cruz"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of California , Santa Cruz","institution_ids":["https://openalex.org/I185103710"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I185103710"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":1,"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":0.8213000297546387,"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":0.8213000297546387,"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":0.06849999725818634,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.02410000003874302,"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/chain-of-thought-reasoning","display_name":"chain-of-thought reasoning","score":0.9940000176429749},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9879999756813049},{"id":"https://openalex.org/keywords/information-gain","display_name":"information gain","score":0.9110000133514404},{"id":"https://openalex.org/keywords/information-theory","display_name":"information theory","score":0.722000002861023},{"id":"https://openalex.org/keywords/failure-mode-identification","display_name":"failure mode identification","score":0.640999972820282},{"id":"https://openalex.org/keywords/arithmetic-reasoning","display_name":"arithmetic reasoning","score":0.8159999847412109},{"id":"https://openalex.org/keywords/gsm8k","display_name":"GSM8K","score":0.9340000152587891}],"concepts":[{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.3746769428253174},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.32970547676086426},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.1464192271232605}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2411.11984","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.11984","pdf_url":"https://arxiv.org/pdf/2411.11984","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.2411.11984","doi":"https://doi.org/10.48550/arxiv.2411.11984","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2411.11984","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:2411.11984","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.11984","pdf_url":"https://arxiv.org/pdf/2411.11984","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":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4404573447.pdf","grobid_xml":"https://content.openalex.org/works/W4404573447.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7127290131","https://openalex.org/W7172495764","https://openalex.org/W4312090934","https://openalex.org/W4404781632","https://openalex.org/W4387226071","https://openalex.org/W7102624217","https://openalex.org/W4403556106","https://openalex.org/W7140106720","https://openalex.org/W4399837031","https://openalex.org/W7154924266"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4],"shown":[5],"impressive":[6],"performance":[7,123],"in":[8,41,61,83],"complex":[9],"reasoning":[10,45,60,75],"tasks":[11],"through":[12,63,99],"the":[13,71,78,86,94],"use":[14],"of":[15,51,80,96],"Chain-of-Thought":[16],"(CoT)":[17],"reasoning,":[18],"allowing":[19],"models":[20],"to":[21,48],"break":[22],"down":[23],"problems":[24],"into":[25,121],"manageable":[26],"sub-tasks.":[27],"However,":[28],"existing":[29,113],"CoT":[30,36,59],"evaluation":[31],"techniques":[32],"either":[33],"require":[34],"annotated":[35,90],"data":[37],"or":[38],"fall":[39],"short":[40],"accurately":[42],"assessing":[43],"intermediate":[44],"steps,":[46],"leading":[47],"high":[49],"rates":[50],"false":[52],"positives.":[53],"In":[54],"this":[55],"paper,":[56],"we":[57],"formalize":[58],"LLMs":[62,84],"an":[64],"information-theoretic":[65],"lens.":[66],"Specifically,":[67],"our":[68,97],"framework":[69],"quantifies":[70],"`information-gain'":[72],"at":[73],"each":[74],"step,":[76],"enabling":[77],"identification":[79],"failure":[81],"modes":[82],"without":[85],"need":[87],"for":[88],"expensive":[89],"datasets.":[91],"We":[92],"demonstrate":[93],"efficacy":[95],"approach":[98],"extensive":[100],"experiments":[101],"on":[102,124],"toy":[103],"arithmetic,":[104],"GSM8K":[105],"and":[106],"PRM800k":[107],"datasets,":[108],"where":[109],"it":[110],"significantly":[111],"outperforms":[112],"outcome-based":[114],"methods":[115],"by":[116],"providing":[117],"more":[118],"accurate":[119],"insights":[120],"model":[122],"individual":[125],"subtasks.":[126]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2024-11-21T00:00:00"}
