{"work":{"id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","openalex_id":"https://openalex.org/W4224308101","doi":"10.48550/arxiv.2204.02311","arxiv_id":"2204.02311","raw_key":null,"title":"PaLM: Scaling Language Modeling with Pathways","authors":null,"authors_text":"Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts","year":2022,"venue":"cs.CL","abstract":"Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pods. We demonstrate continued benefits of scaling by achieving state-of-the-art few-shot learning results on hundreds of language understanding and generation benchmarks. On a number of these tasks, PaLM 540B achieves breakthrough performance, outperforming the finetuned state-of-the-art on a suite of multi-step reasoning tasks, and outperforming average human performance on the recently released BIG-bench benchmark. A significant number of BIG-bench tasks showed discontinuous improvements from model scale, meaning that performance steeply increased as we scaled to our largest model. PaLM also has strong capabilities in multilingual tasks and source code generation, which we demonstrate on a wide array of benchmarks. We additionally provide a comprehensive analysis on bias and toxicity, and study the extent of training data memorization with respect to model scale. Finally, we discuss the ethical considerations related to large language models and discuss potential mitigation strategies.","external_url":"https://arxiv.org/abs/2204.02311","cited_by_count":2131,"metadata_source":"pith","metadata_fetched_at":"2026-08-05T02:28:24.338817+00:00","pith_arxiv_id":"2204.02311","created_at":"2026-05-08T17:13:38.717343+00:00","updated_at":"2026-08-05T02:28:24.338817+00:00","title_quality_ok":true,"display_title":"PaLM: Scaling Language Modeling with Pathways","render_title":"PaLM: Scaling Language Modeling with Pathways"},"hub":{"state":{"work_id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","tier":"super_hub","tier_reason":"100+ Pith inbound or 10,000+ external citations","pith_inbound_count":198,"external_cited_by_count":2131,"distinct_field_count":17,"first_pith_cited_at":"2022-04-01T17:43:13+00:00","last_pith_cited_at":"2026-07-09T06:56:20+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":"2026-08-23T16:19:30.036093+00:00","tier_text":"super_hub"},"tier":"super_hub","role_counts":[{"context_role":"background","n":54},{"context_role":"method","n":2},{"context_role":"other","n":2},{"context_role":"baseline","n":1}],"polarity_counts":[{"context_polarity":"background","n":49},{"context_polarity":"unclear","n":5},{"context_polarity":"support","n":2},{"context_polarity":"use_method","n":2},{"context_polarity":"baseline","n":1}],"runs":{"ask_index":{"job_type":"ask_index","status":"succeeded","result":{"title":"PaLM: Scaling Language Modeling with Pathways","claims":[{"claim_text":"Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pod","claim_type":"abstract","evidence_strength":"source_metadata"},{"claim_text":"Weakly supervised data augmentation through prompting for dialogue understanding. NeurIPS 2022 Workshop on Synthetic Data for Empowering ML Research, 2022. [21] Wei-Lin Chiang, Zhuohan Li, Zi Lin, Ying Sheng, Zhanghao Wu, Hao Zhang, Lianmin Zheng, Siyuan Zhuang, Yonghao Zhuang, Joseph E. Gonzalez, Ion Stoica, and Eric P. Xing. Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality, March 2023. [22] Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra,","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"of latency compared to the state-of-the-art systems, such as FasterTransformer and Orca. The improvement is more pronounced with longer sequences, larger models, and more complex decoding algorithms. vLLM's source code is publicly available at https://github.com/vllm-project/vllm. 1 Introduction The emergence of large language models (LLMs) like GPT [5, 37] and PaLM [9] have enabled new applications such as pro- gramming assistants [6, 18] and universal chatbots [19, 35] that are starting to pro","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"motion generation [39, 6, 32, 61, 5]. However, despite these advancements, current models are mostly task-specific and usually fall short in addressing a broader range of tasks. 2.2. Unified Models Unified models have been developed to address multi- ple tasks. In the language field, Large language models (LLMs) such as GPT-3 [13], LaMDA [57], and PaLM [11] are examples of general-purpose unified models, which han- dle language tasks through an auto-regressive and genera- tive approach. Unlike l","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"tially supported by Wu Tsai Human Performance Alliance Fellowship. References [1] T. Brown, B. Mann, N. Ryder, M. Subbiah, J. D. Kaplan, P. Dhariwal, A. Neelakantan, P. Shyam, G. Sastry, A. Askell, et al. Language models are few-shot learners. Advances in neural information processing systems, 33:1877-1901, 2020. [2] OpenAI. Gpt-4 technical report. arXiv, 2023. [3] A. Chowdhery, S. Narang, J. Devlin, M. Bosma, G. Mishra, A. Roberts, P. Barham, H. W. Chung, C. Sutton, S. Gehrmann, et al. Palm: Sc","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"One of the remarkable observations in recent machine learning research is that large language models (LLMs) can not only generate fluent textual descriptions, but also appear to have rich internalized knowledge about the world [3, 4, 5, 6, 7]. When appropriately conditioned (e.g., prompted), they can even carry out some degree of deduction and respond to questions that appear to require reasoning and inference [8, 9, 10, 11, 12, 13]. This raises an intriguing possibility: beyond their ability to","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"training compute (C), for neural language models. Given a compute budgetc, they empirically presented three basic formulas for the scaling law 6: L(N) = \u0012 Nc N \u0013αN , α N ∼0.076, N c ∼8.8×10 13 (1) L(D) = \u0012 Dc D \u0013αD , α D ∼0.095, D c ∼5.4×10 13 L(C) = \u0012 Cc C \u0013αC , α C ∼0.050, C c ∼3.1×10 8 whereL(·)denotes the cross entropy loss in nats, and a follow-up study [58] from OpenAI has shown that the language modeling loss can be decomposed into two parts, namelyirreducible loss(the entropy of the true","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"}],"why_cited":"Pith tracks PaLM: Scaling Language Modeling with Pathways because it crossed a citation-hub threshold. Current citing contexts most often use it as background evidence (16 contexts).","role_counts":[{"n":16,"context_role":"background"},{"n":1,"context_role":"baseline"},{"n":1,"context_role":"method"}]},"error":null,"updated_at":"2026-05-15T17:48:14.589276+00:00"},"author_expand":{"job_type":"author_expand","status":"succeeded","result":{"authors_linked":[{"id":"d02e9bea-d45d-4246-8e52-818baefa7a0b","orcid":null,"display_name":"Aakanksha Chowdhery"},{"id":"e0e3c462-ade2-43d1-974e-1a346b19be3b","orcid":null,"display_name":"Sharan Narang"},{"id":"22085fc7-7abb-4659-9689-7df10e6d3927","orcid":null,"display_name":"Jacob Devlin"},{"id":"0778e3ce-2eaf-448e-8c23-5a30ba22c9fb","orcid":null,"display_name":"Maarten Bosma"},{"id":"c44b188a-684f-422c-a0ed-cb489ceb32bc","orcid":null,"display_name":"Gaurav Mishra"},{"id":"87b9d64c-539b-4973-b55f-ecc59c89ebfe","orcid":null,"display_name":"Adam Roberts"}]},"error":null,"updated_at":"2026-05-15T17:48:15.702279+00:00"},"context_extract":{"job_type":"context_extract","status":"succeeded","result":{"enqueued_papers":25},"error":null,"updated_at":"2026-05-14T07:38:17.632227+00:00"},"graph_features":{"job_type":"graph_features","status":"succeeded","result":{"co_cited":[{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":33},{"title":"Scaling Laws for Neural Language Models","work_id":"b7dd8749-9c45-4977-ab9b-64478dce1ae8","shared_citers":26},{"title":"Training Compute-Optimal Large Language Models","work_id":"b2faf28d-86b7-429c-bc42-469458efc246","shared_citers":26},{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":24},{"title":"OPT: Open Pre-trained Transformer Language Models","work_id":"d7ff3b21-1fff-4cf4-952a-4714e3ef2307","shared_citers":23},{"title":"LaMDA: Language Models for Dialog Applications","work_id":"1b66d0a5-f6ae-4332-8025-c662dc64b238","shared_citers":22},{"title":"Chain-of-Thought Prompting Elicits Reasoning in Large Language Models","work_id":"d1cf6693-a082-403c-ada9-dac7b96341f9","shared_citers":19},{"title":"Scaling Language Models: Methods, Analysis & Insights from Training Gopher","work_id":"47ce8be9-e500-407d-af41-ac2d132215eb","shared_citers":19},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":19},{"title":"Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models","work_id":"bb63abb3-0d50-4362-b97c-b5e725b03b39","shared_citers":16},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":16},{"title":"On the Opportunities and Risks of Foundation Models","work_id":"a18039e9-928d-47c9-a836-32656a71bf71","shared_citers":16},{"title":"Scaling Instruction-Finetuned Language Models","work_id":"8405abb1-7558-4fdf-af24-f4c52fa77a06","shared_citers":16},{"title":"Training language models to follow instructions with human feedback","work_id":"52aff42f-4fa9-4fcf-bdb3-1459b9bebf65","shared_citers":16},{"title":"Large Language Models are Zero-Shot Reasoners","work_id":"d9b7eb1a-7165-46ff-9f06-d2f0b9d6f95d","shared_citers":15},{"title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","work_id":"68a5177f-d644-44c1-bd4f-4e5278c22f5d","shared_citers":15},{"title":"Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge","work_id":"28ea1282-d657-4c61-a83c-f1249be6d6b1","shared_citers":15},{"title":"RoBERTa: A Robustly Optimized BERT Pretraining Approach","work_id":"41fe12c4-e538-4890-a244-480650ed3078","shared_citers":14},{"title":"Program Synthesis with Large Language Models","work_id":"fd241a05-03b9-4de2-9588-9d77ce176125","shared_citers":13},{"title":"Sparks of Artificial General Intelligence: Early experiments with GPT-4","work_id":"a23cfe92-7f7c-424b-98d4-b386a83002fb","shared_citers":13},{"title":"BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding","work_id":"ed240a10-5b19-406c-baa5-30803f465785","shared_citers":11},{"title":"BLOOM: A 176B-Parameter Open-Access Multilingual Language Model","work_id":"337ba690-f35d-4154-9450-8edf4bc9f488","shared_citers":11},{"title":"Finetuned Language Models Are Zero-Shot Learners","work_id":"7ed6cdaa-ed67-4db4-aceb-b7e1b0e6e7c4","shared_citers":11},{"title":"Least-to-Most Prompting Enables Complex Reasoning in Large Language Models","work_id":"7e58c111-4666-4996-b5ad-1c8efd433083","shared_citers":11}],"time_series":[{"n":11,"year":2022},{"n":39,"year":2023},{"n":9,"year":2024},{"n":1,"year":2025},{"n":22,"year":2026}],"dependency_candidates":[]},"error":null,"updated_at":"2026-05-14T07:38:00.711341+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":"2026-05-14T07:38:09.391512+00:00"},"role_polarity":{"job_type":"role_polarity","status":"succeeded","result":{"title":"PaLM: Scaling Language Modeling with Pathways","claims":[{"claim_text":"Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pod","claim_type":"abstract","evidence_strength":"source_metadata"},{"claim_text":"Weakly supervised data augmentation through prompting for dialogue understanding. NeurIPS 2022 Workshop on Synthetic Data for Empowering ML Research, 2022. [21] Wei-Lin Chiang, Zhuohan Li, Zi Lin, Ying Sheng, Zhanghao Wu, Hao Zhang, Lianmin Zheng, Siyuan Zhuang, Yonghao Zhuang, Joseph E. Gonzalez, Ion Stoica, and Eric P. Xing. Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality, March 2023. [22] Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra,","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"of latency compared to the state-of-the-art systems, such as FasterTransformer and Orca. The improvement is more pronounced with longer sequences, larger models, and more complex decoding algorithms. vLLM's source code is publicly available at https://github.com/vllm-project/vllm. 1 Introduction The emergence of large language models (LLMs) like GPT [5, 37] and PaLM [9] have enabled new applications such as pro- gramming assistants [6, 18] and universal chatbots [19, 35] that are starting to pro","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"motion generation [39, 6, 32, 61, 5]. However, despite these advancements, current models are mostly task-specific and usually fall short in addressing a broader range of tasks. 2.2. Unified Models Unified models have been developed to address multi- ple tasks. In the language field, Large language models (LLMs) such as GPT-3 [13], LaMDA [57], and PaLM [11] are examples of general-purpose unified models, which han- dle language tasks through an auto-regressive and genera- tive approach. Unlike l","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"tially supported by Wu Tsai Human Performance Alliance Fellowship. References [1] T. Brown, B. Mann, N. Ryder, M. Subbiah, J. D. Kaplan, P. Dhariwal, A. Neelakantan, P. Shyam, G. Sastry, A. Askell, et al. Language models are few-shot learners. Advances in neural information processing systems, 33:1877-1901, 2020. [2] OpenAI. Gpt-4 technical report. arXiv, 2023. [3] A. Chowdhery, S. Narang, J. Devlin, M. Bosma, G. Mishra, A. Roberts, P. Barham, H. W. Chung, C. Sutton, S. Gehrmann, et al. Palm: Sc","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"One of the remarkable observations in recent machine learning research is that large language models (LLMs) can not only generate fluent textual descriptions, but also appear to have rich internalized knowledge about the world [3, 4, 5, 6, 7]. When appropriately conditioned (e.g., prompted), they can even carry out some degree of deduction and respond to questions that appear to require reasoning and inference [8, 9, 10, 11, 12, 13]. This raises an intriguing possibility: beyond their ability to","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"training compute (C), for neural language models. Given a compute budgetc, they empirically presented three basic formulas for the scaling law 6: L(N) = \u0012 Nc N \u0013αN , α N ∼0.076, N c ∼8.8×10 13 (1) L(D) = \u0012 Dc D \u0013αD , α D ∼0.095, D c ∼5.4×10 13 L(C) = \u0012 Cc C \u0013αC , α C ∼0.050, C c ∼3.1×10 8 whereL(·)denotes the cross entropy loss in nats, and a follow-up study [58] from OpenAI has shown that the language modeling loss can be decomposed into two parts, namelyirreducible loss(the entropy of the true","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"}],"why_cited":"Pith tracks PaLM: Scaling Language Modeling with Pathways because it crossed a citation-hub threshold. Current citing contexts most often use it as background evidence (16 contexts).","role_counts":[{"n":16,"context_role":"background"},{"n":1,"context_role":"baseline"},{"n":1,"context_role":"method"}]},"error":null,"updated_at":"2026-05-15T17:48:14.594391+00:00"},"summary_claims":{"job_type":"summary_claims","status":"succeeded","result":{"title":"PaLM: Scaling Language Modeling with Pathways","claims":[{"claim_text":"Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pod","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks PaLM: Scaling Language Modeling with Pathways because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-14T07:38:05.112824+00:00"}},"summary":{"title":"PaLM: Scaling Language Modeling with Pathways","claims":[{"claim_text":"Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pod","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks PaLM: Scaling Language Modeling with Pathways because it crossed a citation-hub threshold.","role_counts":[]},"graph":{"co_cited":[{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":33},{"title":"Scaling Laws for Neural Language Models","work_id":"b7dd8749-9c45-4977-ab9b-64478dce1ae8","shared_citers":26},{"title":"Training Compute-Optimal Large Language Models","work_id":"b2faf28d-86b7-429c-bc42-469458efc246","shared_citers":26},{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":24},{"title":"OPT: Open Pre-trained Transformer Language Models","work_id":"d7ff3b21-1fff-4cf4-952a-4714e3ef2307","shared_citers":23},{"title":"LaMDA: Language Models for Dialog Applications","work_id":"1b66d0a5-f6ae-4332-8025-c662dc64b238","shared_citers":22},{"title":"Chain-of-Thought Prompting Elicits Reasoning in Large Language Models","work_id":"d1cf6693-a082-403c-ada9-dac7b96341f9","shared_citers":19},{"title":"Scaling Language Models: Methods, Analysis & Insights from Training Gopher","work_id":"47ce8be9-e500-407d-af41-ac2d132215eb","shared_citers":19},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":19},{"title":"Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models","work_id":"bb63abb3-0d50-4362-b97c-b5e725b03b39","shared_citers":16},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":16},{"title":"On the Opportunities and Risks of Foundation Models","work_id":"a18039e9-928d-47c9-a836-32656a71bf71","shared_citers":16},{"title":"Scaling Instruction-Finetuned Language Models","work_id":"8405abb1-7558-4fdf-af24-f4c52fa77a06","shared_citers":16},{"title":"Training language models to follow instructions with human feedback","work_id":"52aff42f-4fa9-4fcf-bdb3-1459b9bebf65","shared_citers":16},{"title":"Large Language Models are Zero-Shot Reasoners","work_id":"d9b7eb1a-7165-46ff-9f06-d2f0b9d6f95d","shared_citers":15},{"title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","work_id":"68a5177f-d644-44c1-bd4f-4e5278c22f5d","shared_citers":15},{"title":"Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge","work_id":"28ea1282-d657-4c61-a83c-f1249be6d6b1","shared_citers":15},{"title":"RoBERTa: A Robustly Optimized BERT Pretraining Approach","work_id":"41fe12c4-e538-4890-a244-480650ed3078","shared_citers":14},{"title":"Program Synthesis with Large Language Models","work_id":"fd241a05-03b9-4de2-9588-9d77ce176125","shared_citers":13},{"title":"Sparks of Artificial General Intelligence: Early experiments with GPT-4","work_id":"a23cfe92-7f7c-424b-98d4-b386a83002fb","shared_citers":13},{"title":"BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding","work_id":"ed240a10-5b19-406c-baa5-30803f465785","shared_citers":11},{"title":"BLOOM: A 176B-Parameter Open-Access Multilingual Language Model","work_id":"337ba690-f35d-4154-9450-8edf4bc9f488","shared_citers":11},{"title":"Finetuned Language Models Are Zero-Shot Learners","work_id":"7ed6cdaa-ed67-4db4-aceb-b7e1b0e6e7c4","shared_citers":11},{"title":"Least-to-Most Prompting Enables Complex Reasoning in Large Language Models","work_id":"7e58c111-4666-4996-b5ad-1c8efd433083","shared_citers":11}],"time_series":[{"n":11,"year":2022},{"n":39,"year":2023},{"n":9,"year":2024},{"n":1,"year":2025},{"n":22,"year":2026}],"dependency_candidates":[]},"authors":[{"id":"d02e9bea-d45d-4246-8e52-818baefa7a0b","orcid":null,"display_name":"Aakanksha Chowdhery","source":"manual","import_confidence":0.72},{"id":"87b9d64c-539b-4973-b55f-ecc59c89ebfe","orcid":null,"display_name":"Adam Roberts","source":"manual","import_confidence":0.72},{"id":"c44b188a-684f-422c-a0ed-cb489ceb32bc","orcid":null,"display_name":"Gaurav Mishra","source":"manual","import_confidence":0.72},{"id":"22085fc7-7abb-4659-9689-7df10e6d3927","orcid":null,"display_name":"Jacob Devlin","source":"manual","import_confidence":0.72},{"id":"0778e3ce-2eaf-448e-8c23-5a30ba22c9fb","orcid":null,"display_name":"Maarten Bosma","source":"manual","import_confidence":0.72},{"id":"e0e3c462-ade2-43d1-974e-1a346b19be3b","orcid":null,"display_name":"Sharan Narang","source":"manual","import_confidence":0.72}]}}