{"work":{"id":"7e9c3d6e-d6f7-4763-9ef6-de471506c58f","openalex_id":"https://openalex.org/W4376167329","doi":"10.48550/arxiv.2305.06161","arxiv_id":"2305.06161","raw_key":null,"title":"StarCoder: may the source be with you!","authors":null,"authors_text":"Raymond Li, Loubna Ben Allal, Yangtian Zi, Niklas Muennighoff, Denis Kocetkov, Chenghao Mou","year":2023,"venue":"cs.CL","abstract":"The BigCode community, an open-scientific collaboration working on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder and StarCoderBase: 15.5B parameter models with 8K context length, infilling capabilities and fast large-batch inference enabled by multi-query attention. StarCoderBase is trained on 1 trillion tokens sourced from The Stack, a large collection of permissively licensed GitHub repositories with inspection tools and an opt-out process. We fine-tuned StarCoderBase on 35B Python tokens, resulting in the creation of StarCoder. We perform the most comprehensive evaluation of Code LLMs to date and show that StarCoderBase outperforms every open Code LLM that supports multiple programming languages and matches or outperforms the OpenAI code-cushman-001 model. Furthermore, StarCoder outperforms every model that is fine-tuned on Python, can be prompted to achieve 40\\% pass@1 on HumanEval, and still retains its performance on other programming languages. We take several important steps towards a safe open-access model release, including an improved PII redaction pipeline and a novel attribution tracing tool, and make the StarCoder models publicly available under a more commercially viable version of the Open Responsible AI Model license.","external_url":"https://arxiv.org/abs/2305.06161","cited_by_count":192,"metadata_source":"pith","metadata_fetched_at":"2026-08-05T02:28:24.338817+00:00","pith_arxiv_id":"2305.06161","created_at":"2026-05-09T06:05:35.482031+00:00","updated_at":"2026-08-05T02:28:24.338817+00:00","title_quality_ok":false,"display_title":"StarCoder: may the source be with you!","render_title":"StarCoder: may the source be with you!"},"hub":{"state":{"work_id":"7e9c3d6e-d6f7-4763-9ef6-de471506c58f","tier":"super_hub","tier_reason":"100+ Pith inbound or 10,000+ external citations","pith_inbound_count":112,"external_cited_by_count":192,"distinct_field_count":11,"first_pith_cited_at":"2023-03-20T18:08:50+00:00","last_pith_cited_at":"2026-07-09T00:37:25+00:00","author_build_status":"needed","summary_status":"needed","contexts_status":"needed","graph_status":"needed","ask_index_status":"needed","reader_status":"not_needed","recognition_status":"not_needed","updated_at":"2026-08-23T00:29:25.971782+00:00","tier_text":"super_hub"},"tier":"super_hub","role_counts":[{"context_role":"background","n":29}],"polarity_counts":[{"context_polarity":"background","n":26},{"context_polarity":"unclear","n":2},{"context_polarity":"support","n":1}],"runs":{"ask_index":{"job_type":"ask_index","status":"succeeded","result":{"title":"StarCoder: may the source be with you!","claims":[{"claim_text":"The BigCode community, an open-scientific collaboration working on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder and StarCoderBase: 15.5B parameter models with 8K context length, infilling capabilities and fast large-batch inference enabled by multi-query attention. StarCoderBase is trained on 1 trillion tokens sourced from The Stack, a large collection of permissively licensed GitHub repositories with inspection tools and an opt-out process. We fine-tuned StarCoderBase on 35B Python tokens, resulting in the creation of StarCoder. We perform th","claim_type":"abstract","evidence_strength":"source_metadata"},{"claim_text":"in Natural Language Processing (2021). [66] Leviathan, Y., Kalman, M., and Matias, Y. Fast inference from transformers via speculative decoding. In International Conference on Machine Learning (2023), PMLR, pp. 19274-19286. [67] Li, L. H., Hessel, J., Yu, Y., Ren, X., Chang, K.-W., and Choi, Y. Symbolic chain-of-thought distillation: Small models can also \"think\" step-by-step, 2023. [68] Li, R., Allal, L. B., Zi, Y., Muennighoff, N., Kocetkov, D., Mou, C., Marone, M., Akiki, C., Li, J., Chim, J.","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"Model Organization Size Vocab Context Tokens Date CodeGen [165] Salesforce AI 350M, 2B, 6B, 16B 50K 2048 577.2B 2022-03 CodeGen2 [166] Salesforce AI 1B, 3.7B, 7B, 16B 50K 2048 400B-1.4T 2023-03 SantaCoder [167] BigCode 1.1B 48K 2048 236B 2023-01 StarCoder(Base) [168] BigCode 1B, 3B, 7B, 15.5B 48K 8192 1T 2023-05 StarCoder2 [169] BigCode 3B, 7B, 15B 48K 16K 4T 2024-02 Code LLaMA [170] Meta 7B, 13B, 34B, 70B 31K 16K-100K 500B-1T 2023-08 DeepSeek-Coder [171] DeepSeek 1.3B, 5.7B, 6.7B, 33B 32K 16K 2","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pages 74-81. [60] Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, et al. 2021. Few-shot learning with multilingual language models. arXiv preprint arXiv:2112.10668. [61] Yu-Hsiang Lin, Chian-Yu Chen, Jean Lee, Zirui Li, Yuyan Zhang, Mengzhou Xia, Shruti Rijhwani, Junxian He, Zhisong Zhang, Xuezhe Ma, et al. 2019. C","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"(Section 5), discusses implications and threats to validity (Section 6), and concludes with future directions (Section 7). 2 Related Work 2.1 Code Large Language Models Code intelligence has evolved from encoder and encoder-decoder pre-trained models such as CodeBERT [9], GraphCodeBERT [12], CodeT5 [35], and UniXcoder [11] to large decoder-only code LLMs including Codex [ 4], StarCoder [ 5], CodeLlama [ 29], DeepSeek- Coder [13], and Qwen2.5-Coder [15], with progressively stronger capabilities i","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"KeywordsLLM Safety · Malicious Code Generation · Benchmark Dataset · Multi-Judge Classification · Fleiss Kappa 1 Introduction Evaluations of whether large language models (LLMs) refuse malicious requests have become a central activity in AI safety research. In coding-specialized LLMs, open-weight systems such as Code Llama [1], StarCoder and StarCoder2 [2, 3], DeepSeek-Coder [4], and Qwen2.5-Coder [5] now match or exceed closed-model performance on benchmarks such as HumanEval [6]. In this setti","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"Balaji, S. Jain, W. Saunders, C. Hesse, A. N. Carr, J. Leike, J. Achiam, V. Misra, E. Morikawa, A. Radford, M. Knight, M. Brundage, M. Murati, K. Mayer, P. Welinder, B. McGrew, D. Amodei, S. McCandlish, I. Sutskever, W. Zaremba, 28 Evaluating large language models trained on code, arXiv preprint arXiv:2107.03374, 2021. doi:10.48550/arXiv.2107.03374 [2] R. Li, L. B. Allal, Y. Zi, N. Muennighoff, D. Kocetkov, C. Mou, M. Marone, C. Akiki, J. Li, J. Chim, Q. Liu, E. Zheltonozhskii, T. Y. Zhuo, T. Wa","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"}],"why_cited":"Pith tracks StarCoder: may the source be with you! because it crossed a citation-hub threshold. Current citing contexts most often use it as background evidence (29 contexts).","role_counts":[{"n":29,"context_role":"background"}]},"error":null,"updated_at":"2026-07-02T14:02:38.759356+00:00"},"author_expand":{"job_type":"author_expand","status":"succeeded","result":{"authors_linked":[{"id":"56ebc63a-1caf-4898-8ffc-0b055fe7e30b","orcid":null,"display_name":"Raymond Li"},{"id":"f7f3472a-57ff-4f85-b89a-43833a0fd18e","orcid":null,"display_name":"Loubna Ben Allal"},{"id":"1f4a0ba9-5e79-442f-8e22-18142a65f28a","orcid":null,"display_name":"Yangtian Zi"},{"id":"b2b763c9-3a20-46f9-9ecd-d899015c8b0c","orcid":null,"display_name":"Niklas Muennighoff"},{"id":"f1303503-6ca5-4966-90e1-84b7b35178c6","orcid":null,"display_name":"Denis Kocetkov"},{"id":"84f7bd70-82d5-43fd-b665-2f830e7bd0ec","orcid":null,"display_name":"Chenghao Mou"}]},"error":null,"updated_at":"2026-07-02T14:02:39.356118+00:00"},"context_extract":{"job_type":"context_extract","status":"succeeded","result":{"enqueued_papers":25},"error":null,"updated_at":"2026-05-14T17:49:48.762577+00:00"},"graph_features":{"job_type":"graph_features","status":"succeeded","result":{"co_cited":[{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":31},{"title":"Code Llama: Open Foundation Models for Code","work_id":"e73bffa4-7620-47ac-9327-259a60db52ca","shared_citers":23},{"title":"Program Synthesis with Large Language Models","work_id":"fd241a05-03b9-4de2-9588-9d77ce176125","shared_citers":18},{"title":"DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence","work_id":"f22dae5a-27e2-41d0-a061-c4286418dee3","shared_citers":14},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":13},{"title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","work_id":"68a5177f-d644-44c1-bd4f-4e5278c22f5d","shared_citers":12},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":12},{"title":"CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis","work_id":"892a192d-205a-4079-a807-00d7874b392e","shared_citers":9},{"title":"Santacoder: don’t reach for the stars! arXiv preprint arXiv:2301.03988","work_id":"bf393c50-a11b-4a0f-8513-52428ede71f7","shared_citers":9},{"title":"PaLM: Scaling Language Modeling with Pathways","work_id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","shared_citers":8},{"title":"Self-Instruct: Aligning Language Models with Self-Generated Instructions","work_id":"d0018767-775d-406e-861d-539ed681ff73","shared_citers":8},{"title":"StarCoder 2 and The Stack v2: The Next Generation","work_id":"2495cc72-f326-4c23-8a39-d9a08cf583e4","shared_citers":8},{"title":"Measuring Mathematical Problem Solving With the MATH Dataset","work_id":"50652ac6-fb7c-4675-a2c2-159c241feb17","shared_citers":7},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":7},{"title":"Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them","work_id":"513eb205-04ca-4722-9a43-a74e8cbe7e85","shared_citers":6},{"title":"Code- gen2: Lessons for training llms on programming and natural languages","work_id":"0a89ab19-3e11-4b95-a487-5347cb219106","shared_citers":6},{"title":"Gemini: A Family of Highly Capable Multimodal Models","work_id":"83f7c85b-3f11-450f-ac0c-64d9745220b2","shared_citers":6},{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":6},{"title":"Mistral 7B","work_id":"eb5e1305-ad11-4875-ad8d-ad8b8f697599","shared_citers":6},{"title":"Qwen Technical Report","work_id":"bb1fd52f-6b2f-437c-9516-37bdf6eb9be8","shared_citers":6},{"title":"Scaling Laws for Neural Language Models","work_id":"b7dd8749-9c45-4977-ab9b-64478dce1ae8","shared_citers":6},{"title":"Textbooks Are All You Need","work_id":"9b14eca2-9e41-4755-88ac-c3e7b67253f5","shared_citers":6},{"title":"Training Compute-Optimal Large Language Models","work_id":"b2faf28d-86b7-429c-bc42-469458efc246","shared_citers":6},{"title":"arXiv preprint arXiv:2109.00859 , year=","work_id":"23c48a40-bc03-438f-b646-a6c8e3058782","shared_citers":5}],"time_series":[{"n":7,"year":2023},{"n":7,"year":2024},{"n":1,"year":2025},{"n":26,"year":2026}],"dependency_candidates":[]},"error":null,"updated_at":"2026-05-14T17:49:52.561132+00:00"},"identity_refresh":{"job_type":"identity_refresh","status":"succeeded","result":{"items":[{"title":"Qwen3 Technical Report","outcome":"unchanged","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","resolver":"local_arxiv","confidence":0.98,"old_work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e"}],"counts":{"fixed":0,"merged":0,"unchanged":1,"quarantined":0,"needs_external_resolution":0},"errors":[],"attempted":1},"error":null,"updated_at":"2026-05-14T17:49:38.776714+00:00"},"role_polarity":{"job_type":"role_polarity","status":"succeeded","result":{"title":"StarCoder: may the source be with you!","claims":[{"claim_text":"The BigCode community, an open-scientific collaboration working on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder and StarCoderBase: 15.5B parameter models with 8K context length, infilling capabilities and fast large-batch inference enabled by multi-query attention. StarCoderBase is trained on 1 trillion tokens sourced from The Stack, a large collection of permissively licensed GitHub repositories with inspection tools and an opt-out process. We fine-tuned StarCoderBase on 35B Python tokens, resulting in the creation of StarCoder. We perform th","claim_type":"abstract","evidence_strength":"source_metadata"},{"claim_text":"in Natural Language Processing (2021). [66] Leviathan, Y., Kalman, M., and Matias, Y. Fast inference from transformers via speculative decoding. In International Conference on Machine Learning (2023), PMLR, pp. 19274-19286. [67] Li, L. H., Hessel, J., Yu, Y., Ren, X., Chang, K.-W., and Choi, Y. Symbolic chain-of-thought distillation: Small models can also \"think\" step-by-step, 2023. [68] Li, R., Allal, L. B., Zi, Y., Muennighoff, N., Kocetkov, D., Mou, C., Marone, M., Akiki, C., Li, J., Chim, J.","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"Model Organization Size Vocab Context Tokens Date CodeGen [165] Salesforce AI 350M, 2B, 6B, 16B 50K 2048 577.2B 2022-03 CodeGen2 [166] Salesforce AI 1B, 3.7B, 7B, 16B 50K 2048 400B-1.4T 2023-03 SantaCoder [167] BigCode 1.1B 48K 2048 236B 2023-01 StarCoder(Base) [168] BigCode 1B, 3B, 7B, 15.5B 48K 8192 1T 2023-05 StarCoder2 [169] BigCode 3B, 7B, 15B 48K 16K 4T 2024-02 Code LLaMA [170] Meta 7B, 13B, 34B, 70B 31K 16K-100K 500B-1T 2023-08 DeepSeek-Coder [171] DeepSeek 1.3B, 5.7B, 6.7B, 33B 32K 16K 2","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pages 74-81. [60] Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, et al. 2021. Few-shot learning with multilingual language models. arXiv preprint arXiv:2112.10668. [61] Yu-Hsiang Lin, Chian-Yu Chen, Jean Lee, Zirui Li, Yuyan Zhang, Mengzhou Xia, Shruti Rijhwani, Junxian He, Zhisong Zhang, Xuezhe Ma, et al. 2019. C","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"(Section 5), discusses implications and threats to validity (Section 6), and concludes with future directions (Section 7). 2 Related Work 2.1 Code Large Language Models Code intelligence has evolved from encoder and encoder-decoder pre-trained models such as CodeBERT [9], GraphCodeBERT [12], CodeT5 [35], and UniXcoder [11] to large decoder-only code LLMs including Codex [ 4], StarCoder [ 5], CodeLlama [ 29], DeepSeek- Coder [13], and Qwen2.5-Coder [15], with progressively stronger capabilities i","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"KeywordsLLM Safety · Malicious Code Generation · Benchmark Dataset · Multi-Judge Classification · Fleiss Kappa 1 Introduction Evaluations of whether large language models (LLMs) refuse malicious requests have become a central activity in AI safety research. In coding-specialized LLMs, open-weight systems such as Code Llama [1], StarCoder and StarCoder2 [2, 3], DeepSeek-Coder [4], and Qwen2.5-Coder [5] now match or exceed closed-model performance on benchmarks such as HumanEval [6]. In this setti","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"Balaji, S. Jain, W. Saunders, C. Hesse, A. N. Carr, J. Leike, J. Achiam, V. Misra, E. Morikawa, A. Radford, M. Knight, M. Brundage, M. Murati, K. Mayer, P. Welinder, B. McGrew, D. Amodei, S. McCandlish, I. Sutskever, W. Zaremba, 28 Evaluating large language models trained on code, arXiv preprint arXiv:2107.03374, 2021. doi:10.48550/arXiv.2107.03374 [2] R. Li, L. B. Allal, Y. Zi, N. Muennighoff, D. Kocetkov, C. Mou, M. Marone, C. Akiki, J. Li, J. Chim, Q. Liu, E. Zheltonozhskii, T. Y. Zhuo, T. Wa","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"}],"why_cited":"Pith tracks StarCoder: may the source be with you! because it crossed a citation-hub threshold. Current citing contexts most often use it as background evidence (29 contexts).","role_counts":[{"n":29,"context_role":"background"}]},"error":null,"updated_at":"2026-07-02T14:02:38.756893+00:00"},"summary_claims":{"job_type":"summary_claims","status":"succeeded","result":{"title":"StarCoder: may the source be with you!","claims":[{"claim_text":"The BigCode community, an open-scientific collaboration working on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder and StarCoderBase: 15.5B parameter models with 8K context length, infilling capabilities and fast large-batch inference enabled by multi-query attention. StarCoderBase is trained on 1 trillion tokens sourced from The Stack, a large collection of permissively licensed GitHub repositories with inspection tools and an opt-out process. We fine-tuned StarCoderBase on 35B Python tokens, resulting in the creation of StarCoder. We perform th","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks StarCoder: may the source be with you! because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-14T17:49:23.248395+00:00"}},"summary":{"title":"StarCoder: may the source be with you!","claims":[{"claim_text":"The BigCode community, an open-scientific collaboration working on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder and StarCoderBase: 15.5B parameter models with 8K context length, infilling capabilities and fast large-batch inference enabled by multi-query attention. StarCoderBase is trained on 1 trillion tokens sourced from The Stack, a large collection of permissively licensed GitHub repositories with inspection tools and an opt-out process. We fine-tuned StarCoderBase on 35B Python tokens, resulting in the creation of StarCoder. We perform th","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks StarCoder: may the source be with you! because it crossed a citation-hub threshold.","role_counts":[]},"graph":{"co_cited":[{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":31},{"title":"Code Llama: Open Foundation Models for Code","work_id":"e73bffa4-7620-47ac-9327-259a60db52ca","shared_citers":23},{"title":"Program Synthesis with Large Language Models","work_id":"fd241a05-03b9-4de2-9588-9d77ce176125","shared_citers":18},{"title":"DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence","work_id":"f22dae5a-27e2-41d0-a061-c4286418dee3","shared_citers":14},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":13},{"title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","work_id":"68a5177f-d644-44c1-bd4f-4e5278c22f5d","shared_citers":12},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":12},{"title":"CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis","work_id":"892a192d-205a-4079-a807-00d7874b392e","shared_citers":9},{"title":"Santacoder: don’t reach for the stars! arXiv preprint arXiv:2301.03988","work_id":"bf393c50-a11b-4a0f-8513-52428ede71f7","shared_citers":9},{"title":"PaLM: Scaling Language Modeling with Pathways","work_id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","shared_citers":8},{"title":"Self-Instruct: Aligning Language Models with Self-Generated Instructions","work_id":"d0018767-775d-406e-861d-539ed681ff73","shared_citers":8},{"title":"StarCoder 2 and The Stack v2: The Next Generation","work_id":"2495cc72-f326-4c23-8a39-d9a08cf583e4","shared_citers":8},{"title":"Measuring Mathematical Problem Solving With the MATH Dataset","work_id":"50652ac6-fb7c-4675-a2c2-159c241feb17","shared_citers":7},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":7},{"title":"Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them","work_id":"513eb205-04ca-4722-9a43-a74e8cbe7e85","shared_citers":6},{"title":"Code- gen2: Lessons for training llms on programming and natural languages","work_id":"0a89ab19-3e11-4b95-a487-5347cb219106","shared_citers":6},{"title":"Gemini: A Family of Highly Capable Multimodal Models","work_id":"83f7c85b-3f11-450f-ac0c-64d9745220b2","shared_citers":6},{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":6},{"title":"Mistral 7B","work_id":"eb5e1305-ad11-4875-ad8d-ad8b8f697599","shared_citers":6},{"title":"Qwen Technical Report","work_id":"bb1fd52f-6b2f-437c-9516-37bdf6eb9be8","shared_citers":6},{"title":"Scaling Laws for Neural Language Models","work_id":"b7dd8749-9c45-4977-ab9b-64478dce1ae8","shared_citers":6},{"title":"Textbooks Are All You Need","work_id":"9b14eca2-9e41-4755-88ac-c3e7b67253f5","shared_citers":6},{"title":"Training Compute-Optimal Large Language Models","work_id":"b2faf28d-86b7-429c-bc42-469458efc246","shared_citers":6},{"title":"arXiv preprint arXiv:2109.00859 , year=","work_id":"23c48a40-bc03-438f-b646-a6c8e3058782","shared_citers":5}],"time_series":[{"n":7,"year":2023},{"n":7,"year":2024},{"n":1,"year":2025},{"n":26,"year":2026}],"dependency_candidates":[]},"authors":[{"id":"84f7bd70-82d5-43fd-b665-2f830e7bd0ec","orcid":null,"display_name":"Chenghao Mou","source":"manual","import_confidence":0.72},{"id":"f1303503-6ca5-4966-90e1-84b7b35178c6","orcid":null,"display_name":"Denis Kocetkov","source":"manual","import_confidence":0.72},{"id":"f7f3472a-57ff-4f85-b89a-43833a0fd18e","orcid":null,"display_name":"Loubna Ben Allal","source":"manual","import_confidence":0.72},{"id":"b2b763c9-3a20-46f9-9ecd-d899015c8b0c","orcid":null,"display_name":"Niklas Muennighoff","source":"manual","import_confidence":0.72},{"id":"56ebc63a-1caf-4898-8ffc-0b055fe7e30b","orcid":null,"display_name":"Raymond Li","source":"manual","import_confidence":0.72},{"id":"1f4a0ba9-5e79-442f-8e22-18142a65f28a","orcid":null,"display_name":"Yangtian Zi","source":"manual","import_confidence":0.72}]}}