{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:OJEIIRC6PPGPHOZ4JDA3H4YF3B","short_pith_number":"pith:OJEIIRC6","schema_version":"1.0","canonical_sha256":"724884445e7bccf3bb3c48c1b3f305d846cca4ee16af50f662152f0aa95ae36c","source":{"kind":"arxiv","id":"2305.10403","version":3},"attestation_state":"computed","paper":{"title":"PaLM 2 Technical Report","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"PaLM 2 raises quality on English, multilingual, and reasoning tasks while cutting inference time and compute compared to PaLM.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aakanksha Chowdhery, Abe Ittycheriah, Alexandre Passos, Alex Castro Ros, Alex Polozov, Alicia Parrish, Ambrose Slone, Andrea Hu, Andrew M. Dai, Andrew Nystrom, Aroma Mahendru, Aurko Roy, Benjamin Lee, Brennan Saeta, Bryan Richter, Ce Zheng, Chang Lan, Christopher A. Choquette-Choo, Cl\\'ement Crepy, Colin Cherry, Daniel Smilkov, Daniel Sohn, Dasha Valter, David R. So, Denny Zhou, Dmitry Lepikhin, Emanuel Taropa, Emily Reif, Erica Moreira, Eric Chu, Eric Li, Eric Ni, Ethan Dyer, Fangxiaoyu Feng, Frederick Liu, Gaurav Mishra, Gustavo Hernandez Abrego, Guy Gur-Ari, Hadi Hashemi, Hanzhao Lin, Hyeontaek Lim, Jacob Austin, Jacob Devlin, James Bradbury, Jan Botha, Jeffrey Hui, Jeremy Hurwitz, Jiahui Yu, Jian Li, John Nham, John Wieting, Jonathan H. Clark, Joshua Howland, Joshua Maynez, Junwhan Ahn, Katherine Lee, Kathleen Kenealy, Kathy Meier-Hellstern, Kefan Xiao, Kelvin Xu, Kevin Brooks, Kevin Robinson, Kiran Vodrahalli, Laurent El Shafey, Le Hou, Linting Xue, Lucas Gonzalez, Marcello Maggioni, Marie Pellat, Mark D\\'iaz, Mark Omernick, Markus Freitag, Martin Polacek, Matthew Jagielski, Maxim Krikun, Maysam Moussalem, Melvin Johnson, Michael Isard, Michele Catasta, Mostafa Dehghani, Music Li, Nan Du, Orhan Firat, Paige Bailey, Parker Riley, Paul Barham, Pengcheng Yin, Pidong Wang, Qiao Zhang, Rajkumar Samuel, Reiner Pope, Renee Shelby, Rohan Anil, Sebastian Gehrmann, Sebastian Ruder, Shachi Dave, Siamak Shakeri, Siddhartha Brahma, Simon Tokumine, Siyuan Qiao, Slav Petrov, Sneha Kudugunta, Steven Hand, Steven Zheng, Sunipa Dev, Tao Wang, Vedant Misra, Vijay Vasudevan, Vlad Feinberg, Vlad Fienber, Weikang Zhou, Wei Li, Wenhao Jia, Xavier Garcia, Xuezhi Wang, Yaguang Li, Yanping Huang, Yi Tay, Yong Cheng, Yonghui Wu, Yuanzhong Xu, Yuhuai Wu, Yujing Zhang, Yunhan Xu, Zachary Nado, Zhifeng Chen, Zhongtao Liu, Zirui Wang","submitted_at":"2023-05-17T17:46:53Z","abstract_excerpt":"We introduce PaLM 2, a new state-of-the-art language model that has better multilingual and reasoning capabilities and is more compute-efficient than its predecessor PaLM. PaLM 2 is a Transformer-based model trained using a mixture of objectives. Through extensive evaluations on English and multilingual language, and reasoning tasks, we demonstrate that PaLM 2 has significantly improved quality on downstream tasks across different model sizes, while simultaneously exhibiting faster and more efficient inference compared to PaLM. This improved efficiency enables broader deployment while also all"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2305.10403","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-17T17:46:53Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ac37805c5ff5e690c3c8d110c47289e3ea3088176ac37d5f2c326d2978887f55","abstract_canon_sha256":"3a1c4f2d7b48b4ab4b1e818d837be35a2026c4120ade3fe45234bd9c33f7d6d5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:50:29.521922Z","signature_b64":"PwN4NVCZ78Iuq8hVUKjSASYiS13CbmHPDtAQKHEqA/Gh//bKcJwe7FtVoARl05zsHtn5Y199oo8tYsuIQEIMDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"724884445e7bccf3bb3c48c1b3f305d846cca4ee16af50f662152f0aa95ae36c","last_reissued_at":"2026-07-05T06:50:29.521454Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:50:29.521454Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PaLM 2 Technical Report","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"PaLM 2 raises quality on English, multilingual, and reasoning tasks while cutting inference time and compute compared to PaLM.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aakanksha Chowdhery, Abe Ittycheriah, Alexandre Passos, Alex Castro Ros, Alex Polozov, Alicia Parrish, Ambrose Slone, Andrea Hu, Andrew M. Dai, Andrew Nystrom, Aroma Mahendru, Aurko Roy, Benjamin Lee, Brennan Saeta, Bryan Richter, Ce Zheng, Chang Lan, Christopher A. Choquette-Choo, Cl\\'ement Crepy, Colin Cherry, Daniel Smilkov, Daniel Sohn, Dasha Valter, David R. So, Denny Zhou, Dmitry Lepikhin, Emanuel Taropa, Emily Reif, Erica Moreira, Eric Chu, Eric Li, Eric Ni, Ethan Dyer, Fangxiaoyu Feng, Frederick Liu, Gaurav Mishra, Gustavo Hernandez Abrego, Guy Gur-Ari, Hadi Hashemi, Hanzhao Lin, Hyeontaek Lim, Jacob Austin, Jacob Devlin, James Bradbury, Jan Botha, Jeffrey Hui, Jeremy Hurwitz, Jiahui Yu, Jian Li, John Nham, John Wieting, Jonathan H. Clark, Joshua Howland, Joshua Maynez, Junwhan Ahn, Katherine Lee, Kathleen Kenealy, Kathy Meier-Hellstern, Kefan Xiao, Kelvin Xu, Kevin Brooks, Kevin Robinson, Kiran Vodrahalli, Laurent El Shafey, Le Hou, Linting Xue, Lucas Gonzalez, Marcello Maggioni, Marie Pellat, Mark D\\'iaz, Mark Omernick, Markus Freitag, Martin Polacek, Matthew Jagielski, Maxim Krikun, Maysam Moussalem, Melvin Johnson, Michael Isard, Michele Catasta, Mostafa Dehghani, Music Li, Nan Du, Orhan Firat, Paige Bailey, Parker Riley, Paul Barham, Pengcheng Yin, Pidong Wang, Qiao Zhang, Rajkumar Samuel, Reiner Pope, Renee Shelby, Rohan Anil, Sebastian Gehrmann, Sebastian Ruder, Shachi Dave, Siamak Shakeri, Siddhartha Brahma, Simon Tokumine, Siyuan Qiao, Slav Petrov, Sneha Kudugunta, Steven Hand, Steven Zheng, Sunipa Dev, Tao Wang, Vedant Misra, Vijay Vasudevan, Vlad Feinberg, Vlad Fienber, Weikang Zhou, Wei Li, Wenhao Jia, Xavier Garcia, Xuezhi Wang, Yaguang Li, Yanping Huang, Yi Tay, Yong Cheng, Yonghui Wu, Yuanzhong Xu, Yuhuai Wu, Yujing Zhang, Yunhan Xu, Zachary Nado, Zhifeng Chen, Zhongtao Liu, Zirui Wang","submitted_at":"2023-05-17T17:46:53Z","abstract_excerpt":"We introduce PaLM 2, a new state-of-the-art language model that has better multilingual and reasoning capabilities and is more compute-efficient than its predecessor PaLM. PaLM 2 is a Transformer-based model trained using a mixture of objectives. Through extensive evaluations on English and multilingual language, and reasoning tasks, we demonstrate that PaLM 2 has significantly improved quality on downstream tasks across different model sizes, while simultaneously exhibiting faster and more efficient inference compared to PaLM. This improved efficiency enables broader deployment while also all"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"PaLM 2 has significantly improved quality on downstream tasks across different model sizes, while simultaneously exhibiting faster and more efficient inference compared to PaLM.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the suite of English, multilingual, and reasoning benchmarks plus the responsible-AI evaluations accurately capture real-world performance and generalization without undisclosed data selection or post-training adjustments.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PaLM 2 reports state-of-the-art results on language, reasoning, and multilingual tasks with improved efficiency over PaLM.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"PaLM 2 raises quality on English, multilingual, and reasoning tasks while cutting inference time and compute compared to PaLM.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d5d8b7642b976355dba84ec6bd36a4e42c674a59bd51b0ff7aa2afe88593e956"},"source":{"id":"2305.10403","kind":"arxiv","version":3},"verdict":{"id":"cd04cf4c-be46-45c2-b491-49d08336ade8","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-12T11:54:22.599574Z","strongest_claim":"PaLM 2 has significantly improved quality on downstream tasks across different model sizes, while simultaneously exhibiting faster and more efficient inference compared to PaLM.","one_line_summary":"PaLM 2 reports state-of-the-art results on language, reasoning, and multilingual tasks with improved efficiency over PaLM.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the suite of English, multilingual, and reasoning benchmarks plus the responsible-AI evaluations accurately capture real-world performance and generalization without undisclosed data selection or post-training adjustments.","pith_extraction_headline":"PaLM 2 raises quality on English, multilingual, and reasoning tasks while cutting inference time and compute compared to PaLM."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2305.10403/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":286,"sample":[{"doi":"","year":2021,"title":"Persistent anti-muslim bias in large language models","work_id":"ad839eaa-9bbb-4003-b0ca-a13ad8f61202","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Akhbardeh, F., Arkhangorodsky, A., Biesialska, M., Bojar, O., Chatterjee, R., Chaudhary, V., Costa-jussa, M. R., Espa \\ n a-Bonet, C., Fan, A., Federmann, C., Freitag, M., Graham, Y., Grundkiewicz, R.","work_id":"1fc00d99-d922-449b-a078-fe19aaf5b2da","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Guide to fair pay","work_id":"e880db41-23d1-4e7c-814e-a452a6e4f7b1","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback","work_id":"a1f2574b-a899-4713-be60-c87ba332656c","ref_index":5,"cited_arxiv_id":"2204.05862","is_internal_anchor":true},{"doi":"","year":2022,"title":"Building machine translation systems for the next thousand languages","work_id":"c55757ac-fd65-4628-8d13-7ce5b6c20ff0","ref_index":6,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":286,"snapshot_sha256":"b72e054d49232ad8a6b2776522213b2f47cd87521ffa6fdcadc6f5b2d573f5c4","internal_anchors":32},"formal_canon":{"evidence_count":2,"snapshot_sha256":"cf5e23620d2f8316a4a53537d98ba292e8d8689d9b1fecbaae8fe9302b4960e7"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2305.10403","created_at":"2026-07-05T06:50:29.521529+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.10403v3","created_at":"2026-07-05T06:50:29.521529+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.10403","created_at":"2026-07-05T06:50:29.521529+00:00"},{"alias_kind":"pith_short_12","alias_value":"OJEIIRC6PPGP","created_at":"2026-07-05T06:50:29.521529+00:00"},{"alias_kind":"pith_short_16","alias_value":"OJEIIRC6PPGPHOZ4","created_at":"2026-07-05T06:50:29.521529+00:00"},{"alias_kind":"pith_short_8","alias_value":"OJEIIRC6","created_at":"2026-07-05T06:50:29.521529+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":111,"internal_anchor_count":111,"sample":[{"citing_arxiv_id":"2607.08011","citing_title":"Beware What You Autocomplete: Forensic Attribution of Backdoored Code Completions","ref_index":57,"is_internal_anchor":true},{"citing_arxiv_id":"2606.26036","citing_title":"Detect, Unlearn, Restore: Defending Text Summarization Models Against Data Poisoning","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2606.25436","citing_title":"Evaluating Japanese Dialect Robustness Across Speech and Text-based Large Language Models","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2606.26454","citing_title":"Data-driven Machine Learning Cannot Reach Symbolic-level Logical Reasoning -- The Limit of the Scaling Law","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":59,"is_internal_anchor":true},{"citing_arxiv_id":"2606.23000","citing_title":"MotionMAR: Multi-scale Auto-Regressive Human Motion Reconstruction from Sparse Observations","ref_index":46,"is_internal_anchor":true},{"citing_arxiv_id":"2606.17030","citing_title":"Qwen-RobotWorld Technical Report: Unifying Embodied World Modeling through Language-Conditioned Video Generation","ref_index":67,"is_internal_anchor":true},{"citing_arxiv_id":"2606.10931","citing_title":"It Takes One to Bias Them All: Breaking Bad with One-Shot GRPO","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2607.00511","citing_title":"Large Language Models for Multi-Lingual Equivalent Mutant Detection: An Extended Empirical Study","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":59,"is_internal_anchor":true},{"citing_arxiv_id":"2606.30989","citing_title":"Wait, am I Being Fair? Characterizing Deductive Stereotyping and Mitigating It with Fair-GCG","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12890","citing_title":"Steer-to-Detect: Probing Hidden Representations for Detection of LLM-Generated Texts","ref_index":67,"is_internal_anchor":true},{"citing_arxiv_id":"2605.24981","citing_title":"Large Language Model Selection with Limited Annotations","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.30311","citing_title":"Archon: A Unified Multimodal Model for Holistic Digital Human Generation","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2602.20102","citing_title":"BarrierSteer: LLM Safety via Learning Barrier Steering","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2309.10305","citing_title":"Baichuan 2: Open Large-scale Language Models","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2310.11453","citing_title":"BitNet: Scaling 1-bit Transformers for Large Language Models","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2311.07911","citing_title":"Instruction-Following Evaluation for Large Language Models","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2403.03952","citing_title":"Bridging Language and Items for Retrieval and Recommendation: Benchmarking LLMs as Semantic Encoders","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2404.07475","citing_title":"Laissez-Faire Harms: Algorithmic Biases in Generative Language Models","ref_index":32,"is_internal_anchor":true},{"citing_arxiv_id":"2405.14093","citing_title":"A Survey on Vision-Language-Action Models for Embodied AI","ref_index":290,"is_internal_anchor":true},{"citing_arxiv_id":"2408.12935","citing_title":"AI Safety Landscape for Large Language Models: Taxonomy, State-of-the-art, and Future Directions","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2410.04960","citing_title":"On Efficient Variants of Segment Anything Model: A Survey","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2410.13181","citing_title":"AdaSwitch: Adaptive Switching between Small and Large Agents for Effective Cloud-Local Collaborative Learning","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2411.10446","citing_title":"VeriGraph: Scene Graphs for Execution Verifiable Robot Planning","ref_index":37,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B","json":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B.json","graph_json":"https://pith.science/api/pith-number/OJEIIRC6PPGPHOZ4JDA3H4YF3B/graph.json","events_json":"https://pith.science/api/pith-number/OJEIIRC6PPGPHOZ4JDA3H4YF3B/events.json","paper":"https://pith.science/paper/OJEIIRC6"},"agent_actions":{"view_html":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B","download_json":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B.json","view_paper":"https://pith.science/paper/OJEIIRC6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.10403&json=true","fetch_graph":"https://pith.science/api/pith-number/OJEIIRC6PPGPHOZ4JDA3H4YF3B/graph.json","fetch_events":"https://pith.science/api/pith-number/OJEIIRC6PPGPHOZ4JDA3H4YF3B/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B/action/storage_attestation","attest_author":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B/action/author_attestation","sign_citation":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B/action/citation_signature","submit_replication":"https://pith.science/pith/OJEIIRC6PPGPHOZ4JDA3H4YF3B/action/replication_record"}},"created_at":"2026-07-05T06:50:29.521529+00:00","updated_at":"2026-07-05T06:50:29.521529+00:00"}