{"created":"2025-01-19T01:37:05.605515+00:00","updated":"2025-01-19T09:36:44.275452+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00235098","sets":["1164:4179:11560:11676"]},"path":["11676"],"owner":"44499","recid":"235098","title":["自動作成した評価観点を用いたLLMによるLLMの参照回答ベース評価"],"pubdate":{"attribute_name":"公開日","attribute_value":"2024-06-21"},"_buckets":{"deposit":"e5ed2dcb-294e-492c-84f8-7aecce978163"},"_deposit":{"id":"235098","pid":{"type":"depid","value":"235098","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"自動作成した評価観点を用いたLLMによるLLMの参照回答ベース評価","author_link":["641865","641866","641864","641867"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"自動作成した評価観点を用いたLLMによるLLMの参照回答ベース評価"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"LLM応用","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2024-06-21","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"同志社大学"},{"subitem_text_value":"トランス・コスモス株式会社"},{"subitem_text_value":"同志社大学"},{"subitem_text_value":"トランス・コスモス株式会社"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Doshisha University","subitem_text_language":"en"},{"subitem_text_value":"transcosmos inc.","subitem_text_language":"en"},{"subitem_text_value":"Doshisha University","subitem_text_language":"en"},{"subitem_text_value":"transcosmos inc.","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/235098/files/IPSJ-NL24260007.pdf","label":"IPSJ-NL24260007.pdf"},"date":[{"dateType":"Available","dateValue":"2026-06-21"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-NL24260007.pdf","filesize":[{"value":"1.1 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"23"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"bca19f7f-9308-4b86-b065-bb51d62d475a","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2024 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"森田, 隼功"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"大林, 弘明"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"田村, 晃裕"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"濱田, 充男"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10115061","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8779","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"本論文では，大規模言語モデル（LLM）による LLM の新たな性能評価手法を提案する．LLM による LLM の性能評価には，各 LLM の生成文を独立に採点する絶対評価と，生成文同士を比較して優劣を評価する相対評価とがある．従来の絶対評価では，各 LLM の評価が異なる視点で行われる可能性があり，LLM 間の点数を比較することが不適切な場合がある．また，従来の相対評価では，評価対象の提示順によって優劣結果が入れ替わる場合がある．これらの問題は，評価対象の LLM を採点もしくは比較する際，同じ視点で評価が行われていないことが原因であると考える．そこで本研究では，LLM により，質問と参照回答を基に評価観点を自動で作成し，評価観点を指定して LLM を評価する手法を提案する．そして，本提案手法を文章作成等のビジネスタスクを対象に評価し，有効性を確認した．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"14","bibliographic_titles":[{"bibliographic_title":"研究報告自然言語処理（NL）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2024-06-21","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"7","bibliographicVolumeNumber":"2024-NL-260"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":235098,"links":{}}