{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:DCDPKVRXTDIYEJSDLF7FJSWYFP","short_pith_number":"pith:DCDPKVRX","schema_version":"1.0","canonical_sha256":"1886f5563798d1822643597e54cad82bc3209fdbb2bc6bc31a5549acfed333f8","source":{"kind":"arxiv","id":"2304.08354","version":3},"attestation_state":"computed","paper":{"title":"Tool Learning with Foundation Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Foundation models learn tool use by decomposing tasks into subtasks, reasoning to adjust plans, and selecting the right tools for each step.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bokai Xu, Bowen Li, Chaojun Xiao, Cheng Qian, Cheng Yang, Chi Han, Dahai Li, Ganqu Cui, Heng Ji, Huadong Wang, Jason Phang, Jing Yi, Junxi Yan, Kunlun Zhu, Lan Yan, Maosong Sun, Ning Ding, Runchu Tian, Shengding Hu, Shihao Liang, Tongshuang Wu, Weilin Zhao, Weize Chen, Xian Sun, Xin Cong, Xingyu Shen, Xu Han, Yankai Lin, Yaxi Lu, Yining Ye, Yi Ren Fung, Yufei Huang, Yujia Qin, Yusheng Su, Yuxiang Huang, Yuzhang Zhu, Zheni Zeng, Zhenning Dai, Zhen Zhang, Zhiyuan Liu, Ziwei Tang","submitted_at":"2023-04-17T15:16:10Z","abstract_excerpt":"Humans possess an extraordinary ability to create and utilize tools, allowing them to overcome physical limitations and explore new frontiers. With the advent of foundation models, AI systems have the potential to be equally adept in tool use as humans. This paradigm, i.e., tool learning with foundation models, combines the strengths of specialized tools and foundation models to achieve enhanced accuracy, efficiency, and automation in problem-solving. Despite its immense potential, there is still a lack of a comprehensive understanding of key challenges, opportunities, and future endeavors in "},"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":"2304.08354","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-04-17T15:16:10Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"4cc0b63cd7974af7923f4c96d284caa7265f6d56a40398ea3cf20a4d8d82977c","abstract_canon_sha256":"b52d902d9c90a4c8075368c709429af6c316cb03eae703881a1705641afc0d71"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:47.771106Z","signature_b64":"wERVRdxRjQWUd71xk3CHAJx9YWicdwVE/zxR+loXJOxPYH2QBwEjQFmsER6Uvo2bv7CM11ibdVXF0vGEKtyKAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1886f5563798d1822643597e54cad82bc3209fdbb2bc6bc31a5549acfed333f8","last_reissued_at":"2026-05-17T23:38:47.770663Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:47.770663Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Tool Learning with Foundation Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Foundation models learn tool use by decomposing tasks into subtasks, reasoning to adjust plans, and selecting the right tools for each step.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bokai Xu, Bowen Li, Chaojun Xiao, Cheng Qian, Cheng Yang, Chi Han, Dahai Li, Ganqu Cui, Heng Ji, Huadong Wang, Jason Phang, Jing Yi, Junxi Yan, Kunlun Zhu, Lan Yan, Maosong Sun, Ning Ding, Runchu Tian, Shengding Hu, Shihao Liang, Tongshuang Wu, Weilin Zhao, Weize Chen, Xian Sun, Xin Cong, Xingyu Shen, Xu Han, Yankai Lin, Yaxi Lu, Yining Ye, Yi Ren Fung, Yufei Huang, Yujia Qin, Yusheng Su, Yuxiang Huang, Yuzhang Zhu, Zheni Zeng, Zhenning Dai, Zhen Zhang, Zhiyuan Liu, Ziwei Tang","submitted_at":"2023-04-17T15:16:10Z","abstract_excerpt":"Humans possess an extraordinary ability to create and utilize tools, allowing them to overcome physical limitations and explore new frontiers. With the advent of foundation models, AI systems have the potential to be equally adept in tool use as humans. This paradigm, i.e., tool learning with foundation models, combines the strengths of specialized tools and foundation models to achieve enhanced accuracy, efficiency, and automation in problem-solving. Despite its immense potential, there is still a lack of a comprehensive understanding of key challenges, opportunities, and future endeavors in "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We formulate a general tool learning framework: starting from understanding the user instruction, models should learn to decompose a complex task into several subtasks, dynamically adjust their plan through reasoning, and effectively conquer each sub-task by selecting appropriate tools.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That foundation models can be effectively trained and prompted to follow the proposed decomposition-reasoning-tool-selection process at scale, and that experiments with 18 tools sufficiently demonstrate this potential without detailed methodology or baselines in the provided abstract.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"The paper reviews tool learning research, formulates a framework for foundation models to decompose tasks and use tools via reasoning, and demonstrates current models' capabilities through experiments with 18 tools.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Foundation models learn tool use by decomposing tasks into subtasks, reasoning to adjust plans, and selecting the right tools for each step.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2412f454c51c015c229cdaef3fa08076e1d5dcf308cdec892680976007bb2803"},"source":{"id":"2304.08354","kind":"arxiv","version":3},"verdict":{"id":"76c29493-1ce1-47ce-afe6-d4bb816605b2","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T13:06:29.151925Z","strongest_claim":"We formulate a general tool learning framework: starting from understanding the user instruction, models should learn to decompose a complex task into several subtasks, dynamically adjust their plan through reasoning, and effectively conquer each sub-task by selecting appropriate tools.","one_line_summary":"The paper reviews tool learning research, formulates a framework for foundation models to decompose tasks and use tools via reasoning, and demonstrates current models' capabilities through experiments with 18 tools.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That foundation models can be effectively trained and prompted to follow the proposed decomposition-reasoning-tool-selection process at scale, and that experiments with 18 tools sufficiently demonstrate this potential without detailed methodology or baselines in the provided abstract.","pith_extraction_headline":"Foundation models learn tool use by decomposing tasks into subtasks, reasoning to adjust plans, and selecting the right tools for each step."},"references":{"count":15,"sample":[{"doi":"","year":1909,"title":"Fine-Tuning Language Models from Human Preferences","work_id":"4f54aad1-f3b6-404f-b9c7-e21ba0a33b99","ref_index":1,"cited_arxiv_id":"1909.08593","is_internal_anchor":true},{"doi":"","year":null,"title":"tag: Banana Pie Recipes, type: recipe","work_id":"abd865bc-8225-4696-acd2-fc6af90a5d5d","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"tag: Custard and Cream Pies, type: recipes","work_id":"2877694b-2a06-413f-931b-047c242b2b73","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"tag: Mexican, type: recipe","work_id":"d40f9801-993b-43de-9ff6-aa34deda402c","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"tag: No-Bake Pie Recipes, type: recipe","work_id":"e2faeb4d-3fca-4306-b929-aef0dad68d4f","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":15,"snapshot_sha256":"24703edc8d3e872b858f72cbee6240e4252a22edc057853df73efb236884d513","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ae5a9e33860fd411e5c291e032aa4599bc12d2abf4dc592ca4250f7429a5eb51"},"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":"2304.08354","created_at":"2026-05-17T23:38:47.770734+00:00"},{"alias_kind":"arxiv_version","alias_value":"2304.08354v3","created_at":"2026-05-17T23:38:47.770734+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.08354","created_at":"2026-05-17T23:38:47.770734+00:00"},{"alias_kind":"pith_short_12","alias_value":"DCDPKVRXTDIY","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_16","alias_value":"DCDPKVRXTDIYEJSD","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_8","alias_value":"DCDPKVRX","created_at":"2026-05-18T12:33:33.725879+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":22,"internal_anchor_count":22,"sample":[{"citing_arxiv_id":"2409.00557","citing_title":"Learning to Ask: When LLM Agents Meet Unclear Instruction","ref_index":15,"is_internal_anchor":true},{"citing_arxiv_id":"2405.07960","citing_title":"AgentClinic: a multimodal agent benchmark to evaluate AI in simulated clinical environments","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2605.18133","citing_title":"An Empirical Study of Privacy Leakage Chains via Prompt Injection in Black-Box Chatbot Environments","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2401.05459","citing_title":"Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security","ref_index":241,"is_internal_anchor":true},{"citing_arxiv_id":"2306.06070","citing_title":"Mind2Web: Towards a Generalist Agent for the Web","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2305.18323","citing_title":"ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Models","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2403.17297","citing_title":"InternLM2 Technical Report","ref_index":155,"is_internal_anchor":true},{"citing_arxiv_id":"2304.15010","citing_title":"LLaMA-Adapter V2: Parameter-Efficient Visual Instruction Model","ref_index":51,"is_internal_anchor":true},{"citing_arxiv_id":"2404.13501","citing_title":"A Survey on the Memory Mechanism of Large Language Model based Agents","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09038","citing_title":"SearchSkill: Teaching LLMs to Use Search Tools with Evolving Skill Banks","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2308.11432","citing_title":"A Survey on Large Language Model based Autonomous Agents","ref_index":151,"is_internal_anchor":true},{"citing_arxiv_id":"2605.15184","citing_title":"Is Grep All You Need? How Agent Harnesses Reshape Agentic Search","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2504.13958","citing_title":"ToolRL: Reward is All Tool Learning Needs","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2402.02716","citing_title":"Understanding the planning of LLM agents: A survey","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2401.10774","citing_title":"Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads","ref_index":91,"is_internal_anchor":true},{"citing_arxiv_id":"2410.23218","citing_title":"OS-ATLAS: A Foundation Action Model for Generalist GUI Agents","ref_index":99,"is_internal_anchor":true},{"citing_arxiv_id":"2309.01219","citing_title":"Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Models","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09038","citing_title":"SearchSkill: Teaching LLMs to Use Search Tools with Evolving Skill Banks","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09544","citing_title":"TIDE-Bench: Task-Aware and Diagnostic Evaluation of Tool-Integrated Reasoning","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00060","citing_title":"TADI: Tool-Augmented Drilling Intelligence via Agentic LLM Orchestration over Heterogeneous Wellsite Data","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2309.07864","citing_title":"The Rise and Potential of Large Language Model Based Agents: A Survey","ref_index":95,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07675","citing_title":"FactoryBench: Evaluating Industrial Machine Understanding","ref_index":31,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP","json":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP.json","graph_json":"https://pith.science/api/pith-number/DCDPKVRXTDIYEJSDLF7FJSWYFP/graph.json","events_json":"https://pith.science/api/pith-number/DCDPKVRXTDIYEJSDLF7FJSWYFP/events.json","paper":"https://pith.science/paper/DCDPKVRX"},"agent_actions":{"view_html":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP","download_json":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP.json","view_paper":"https://pith.science/paper/DCDPKVRX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2304.08354&json=true","fetch_graph":"https://pith.science/api/pith-number/DCDPKVRXTDIYEJSDLF7FJSWYFP/graph.json","fetch_events":"https://pith.science/api/pith-number/DCDPKVRXTDIYEJSDLF7FJSWYFP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP/action/storage_attestation","attest_author":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP/action/author_attestation","sign_citation":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP/action/citation_signature","submit_replication":"https://pith.science/pith/DCDPKVRXTDIYEJSDLF7FJSWYFP/action/replication_record"}},"created_at":"2026-05-17T23:38:47.770734+00:00","updated_at":"2026-05-17T23:38:47.770734+00:00"}