KwaiYiiMath: Technical Report
Recent advancements in large language models (LLMs) have demonstrated remarkable abilities in handling a variety of natural language processing (NLP) downstream tasks, even on mathematical tasks requiring multi-step reasoning. In this report, we introduce the KwaiYiiMath which enhances the mathemati...
Gespeichert in:
| Veröffentlicht in: | arXiv.org 2023-10 |
|---|---|
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | |
|---|---|
| container_issue | |
| container_start_page | |
| container_title | arXiv.org |
| container_volume | |
| creator | Fu, Jiayi Lin, Lei Gao, Xiaoyang Liu, Pengli Chen, Zhengzong Yang, Zhirui Zhang, Shengnan Zheng, Xue Li, Yan Liu, Yuliang Ye, Xucheng Liao, Yiqiao Liao, Chao Chen, Bin Song, Chengru Wan, Junchen Lin, Zijia Zhang, Fuzheng Wang, Zhongyuan Zhang, Di Gai, Kun |
| description | Recent advancements in large language models (LLMs) have demonstrated remarkable abilities in handling a variety of natural language processing (NLP) downstream tasks, even on mathematical tasks requiring multi-step reasoning. In this report, we introduce the KwaiYiiMath which enhances the mathematical reasoning abilities of KwaiYiiBase1, by applying Supervised Fine-Tuning (SFT) and Reinforced Learning from Human Feedback (RLHF), including on both English and Chinese mathematical tasks. Meanwhile, we also constructed a small-scale Chinese primary school mathematics test set (named KMath), consisting of 188 examples to evaluate the correctness of the problem-solving process generated by the models. Empirical studies demonstrate that KwaiYiiMath can achieve state-of-the-art (SOTA) performance on GSM8k, CMath, and KMath compared with the similar size models, respectively. |
| format | Article |
| fullrecord | <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2876195169</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2876195169</sourcerecordid><originalsourceid>FETCH-proquest_journals_28761951693</originalsourceid><addsrcrecordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mSQ9S5PzIzMzPRNLMmwUghJTc7Iy0xOzFEISi3ILyrhYWBNS8wpTuWF0twMym6uIc4eugVF-YWlqcUl8Vn5pUV5QKl4IwtzM0NLU0MzS2PiVAEAQf4r0A</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2876195169</pqid></control><display><type>article</type><title>KwaiYiiMath: Technical Report</title><source>Free E- Journals</source><creator>Fu, Jiayi ; Lin, Lei ; Gao, Xiaoyang ; Liu, Pengli ; Chen, Zhengzong ; Yang, Zhirui ; Zhang, Shengnan ; Zheng, Xue ; Li, Yan ; Liu, Yuliang ; Ye, Xucheng ; Liao, Yiqiao ; Liao, Chao ; Chen, Bin ; Song, Chengru ; Wan, Junchen ; Lin, Zijia ; Zhang, Fuzheng ; Wang, Zhongyuan ; Zhang, Di ; Gai, Kun</creator><creatorcontrib>Fu, Jiayi ; Lin, Lei ; Gao, Xiaoyang ; Liu, Pengli ; Chen, Zhengzong ; Yang, Zhirui ; Zhang, Shengnan ; Zheng, Xue ; Li, Yan ; Liu, Yuliang ; Ye, Xucheng ; Liao, Yiqiao ; Liao, Chao ; Chen, Bin ; Song, Chengru ; Wan, Junchen ; Lin, Zijia ; Zhang, Fuzheng ; Wang, Zhongyuan ; Zhang, Di ; Gai, Kun</creatorcontrib><description>Recent advancements in large language models (LLMs) have demonstrated remarkable abilities in handling a variety of natural language processing (NLP) downstream tasks, even on mathematical tasks requiring multi-step reasoning. In this report, we introduce the KwaiYiiMath which enhances the mathematical reasoning abilities of KwaiYiiBase1, by applying Supervised Fine-Tuning (SFT) and Reinforced Learning from Human Feedback (RLHF), including on both English and Chinese mathematical tasks. Meanwhile, we also constructed a small-scale Chinese primary school mathematics test set (named KMath), consisting of 188 examples to evaluate the correctness of the problem-solving process generated by the models. Empirical studies demonstrate that KwaiYiiMath can achieve state-of-the-art (SOTA) performance on GSM8k, CMath, and KMath compared with the similar size models, respectively.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Large language models ; Mathematical analysis ; Natural language processing ; Reasoning</subject><ispartof>arXiv.org, 2023-10</ispartof><rights>2023. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>780,784</link.rule.ids></links><search><creatorcontrib>Fu, Jiayi</creatorcontrib><creatorcontrib>Lin, Lei</creatorcontrib><creatorcontrib>Gao, Xiaoyang</creatorcontrib><creatorcontrib>Liu, Pengli</creatorcontrib><creatorcontrib>Chen, Zhengzong</creatorcontrib><creatorcontrib>Yang, Zhirui</creatorcontrib><creatorcontrib>Zhang, Shengnan</creatorcontrib><creatorcontrib>Zheng, Xue</creatorcontrib><creatorcontrib>Li, Yan</creatorcontrib><creatorcontrib>Liu, Yuliang</creatorcontrib><creatorcontrib>Ye, Xucheng</creatorcontrib><creatorcontrib>Liao, Yiqiao</creatorcontrib><creatorcontrib>Liao, Chao</creatorcontrib><creatorcontrib>Chen, Bin</creatorcontrib><creatorcontrib>Song, Chengru</creatorcontrib><creatorcontrib>Wan, Junchen</creatorcontrib><creatorcontrib>Lin, Zijia</creatorcontrib><creatorcontrib>Zhang, Fuzheng</creatorcontrib><creatorcontrib>Wang, Zhongyuan</creatorcontrib><creatorcontrib>Zhang, Di</creatorcontrib><creatorcontrib>Gai, Kun</creatorcontrib><title>KwaiYiiMath: Technical Report</title><title>arXiv.org</title><description>Recent advancements in large language models (LLMs) have demonstrated remarkable abilities in handling a variety of natural language processing (NLP) downstream tasks, even on mathematical tasks requiring multi-step reasoning. In this report, we introduce the KwaiYiiMath which enhances the mathematical reasoning abilities of KwaiYiiBase1, by applying Supervised Fine-Tuning (SFT) and Reinforced Learning from Human Feedback (RLHF), including on both English and Chinese mathematical tasks. Meanwhile, we also constructed a small-scale Chinese primary school mathematics test set (named KMath), consisting of 188 examples to evaluate the correctness of the problem-solving process generated by the models. Empirical studies demonstrate that KwaiYiiMath can achieve state-of-the-art (SOTA) performance on GSM8k, CMath, and KMath compared with the similar size models, respectively.</description><subject>Large language models</subject><subject>Mathematical analysis</subject><subject>Natural language processing</subject><subject>Reasoning</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mSQ9S5PzIzMzPRNLMmwUghJTc7Iy0xOzFEISi3ILyrhYWBNS8wpTuWF0twMym6uIc4eugVF-YWlqcUl8Vn5pUV5QKl4IwtzM0NLU0MzS2PiVAEAQf4r0A</recordid><startdate>20231019</startdate><enddate>20231019</enddate><creator>Fu, Jiayi</creator><creator>Lin, Lei</creator><creator>Gao, Xiaoyang</creator><creator>Liu, Pengli</creator><creator>Chen, Zhengzong</creator><creator>Yang, Zhirui</creator><creator>Zhang, Shengnan</creator><creator>Zheng, Xue</creator><creator>Li, Yan</creator><creator>Liu, Yuliang</creator><creator>Ye, Xucheng</creator><creator>Liao, Yiqiao</creator><creator>Liao, Chao</creator><creator>Chen, Bin</creator><creator>Song, Chengru</creator><creator>Wan, Junchen</creator><creator>Lin, Zijia</creator><creator>Zhang, Fuzheng</creator><creator>Wang, Zhongyuan</creator><creator>Zhang, Di</creator><creator>Gai, Kun</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20231019</creationdate><title>KwaiYiiMath: Technical Report</title><author>Fu, Jiayi ; Lin, Lei ; Gao, Xiaoyang ; Liu, Pengli ; Chen, Zhengzong ; Yang, Zhirui ; Zhang, Shengnan ; Zheng, Xue ; Li, Yan ; Liu, Yuliang ; Ye, Xucheng ; Liao, Yiqiao ; Liao, Chao ; Chen, Bin ; Song, Chengru ; Wan, Junchen ; Lin, Zijia ; Zhang, Fuzheng ; Wang, Zhongyuan ; Zhang, Di ; Gai, Kun</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_28761951693</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Large language models</topic><topic>Mathematical analysis</topic><topic>Natural language processing</topic><topic>Reasoning</topic><toplevel>online_resources</toplevel><creatorcontrib>Fu, Jiayi</creatorcontrib><creatorcontrib>Lin, Lei</creatorcontrib><creatorcontrib>Gao, Xiaoyang</creatorcontrib><creatorcontrib>Liu, Pengli</creatorcontrib><creatorcontrib>Chen, Zhengzong</creatorcontrib><creatorcontrib>Yang, Zhirui</creatorcontrib><creatorcontrib>Zhang, Shengnan</creatorcontrib><creatorcontrib>Zheng, Xue</creatorcontrib><creatorcontrib>Li, Yan</creatorcontrib><creatorcontrib>Liu, Yuliang</creatorcontrib><creatorcontrib>Ye, Xucheng</creatorcontrib><creatorcontrib>Liao, Yiqiao</creatorcontrib><creatorcontrib>Liao, Chao</creatorcontrib><creatorcontrib>Chen, Bin</creatorcontrib><creatorcontrib>Song, Chengru</creatorcontrib><creatorcontrib>Wan, Junchen</creatorcontrib><creatorcontrib>Lin, Zijia</creatorcontrib><creatorcontrib>Zhang, Fuzheng</creatorcontrib><creatorcontrib>Wang, Zhongyuan</creatorcontrib><creatorcontrib>Zhang, Di</creatorcontrib><creatorcontrib>Gai, Kun</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Access via ProQuest (Open Access)</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Fu, Jiayi</au><au>Lin, Lei</au><au>Gao, Xiaoyang</au><au>Liu, Pengli</au><au>Chen, Zhengzong</au><au>Yang, Zhirui</au><au>Zhang, Shengnan</au><au>Zheng, Xue</au><au>Li, Yan</au><au>Liu, Yuliang</au><au>Ye, Xucheng</au><au>Liao, Yiqiao</au><au>Liao, Chao</au><au>Chen, Bin</au><au>Song, Chengru</au><au>Wan, Junchen</au><au>Lin, Zijia</au><au>Zhang, Fuzheng</au><au>Wang, Zhongyuan</au><au>Zhang, Di</au><au>Gai, Kun</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>KwaiYiiMath: Technical Report</atitle><jtitle>arXiv.org</jtitle><date>2023-10-19</date><risdate>2023</risdate><eissn>2331-8422</eissn><abstract>Recent advancements in large language models (LLMs) have demonstrated remarkable abilities in handling a variety of natural language processing (NLP) downstream tasks, even on mathematical tasks requiring multi-step reasoning. In this report, we introduce the KwaiYiiMath which enhances the mathematical reasoning abilities of KwaiYiiBase1, by applying Supervised Fine-Tuning (SFT) and Reinforced Learning from Human Feedback (RLHF), including on both English and Chinese mathematical tasks. Meanwhile, we also constructed a small-scale Chinese primary school mathematics test set (named KMath), consisting of 188 examples to evaluate the correctness of the problem-solving process generated by the models. Empirical studies demonstrate that KwaiYiiMath can achieve state-of-the-art (SOTA) performance on GSM8k, CMath, and KMath compared with the similar size models, respectively.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext |
| identifier | EISSN: 2331-8422 |
| ispartof | arXiv.org, 2023-10 |
| issn | 2331-8422 |
| language | eng |
| recordid | cdi_proquest_journals_2876195169 |
| source | Free E- Journals |
| subjects | Large language models Mathematical analysis Natural language processing Reasoning |
| title | KwaiYiiMath: Technical Report |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-01T15%3A40%3A08IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=KwaiYiiMath:%20Technical%20Report&rft.jtitle=arXiv.org&rft.au=Fu,%20Jiayi&rft.date=2023-10-19&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2876195169%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2876195169&rft_id=info:pmid/&rfr_iscdi=true |