{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:P2NJVCTEQWIAE4OECRX2VOAPOU","short_pith_number":"pith:P2NJVCTE","schema_version":"1.0","canonical_sha256":"7e9a9a8a6485900271c4146faab80f7536f491c52d7415ea2ed5f55285921c9b","source":{"kind":"arxiv","id":"2505.16410","version":1},"attestation_state":"computed","paper":{"title":"Tool-Star: Empowering LLM-Brained Multi-Tool Reasoner via Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Guanting Dong, Guorui Zhou, Hangyu Mao, Hongjin Qian, Jiajie Jin, Ji-Rong Wen, Xiaoxi Li, Yifei Chen, Yutao Zhu, Zhicheng Dou","submitted_at":"2025-05-22T09:00:19Z","abstract_excerpt":"Recently, large language models (LLMs) have shown remarkable reasoning capabilities via large-scale reinforcement learning (RL). However, leveraging the RL algorithm to empower effective multi-tool collaborative reasoning in LLMs remains an open challenge. In this paper, we introduce Tool-Star, an RL-based framework designed to empower LLMs to autonomously invoke multiple external tools during stepwise reasoning. Tool-Star integrates six types of tools and incorporates systematic designs in both data synthesis and training. To address the scarcity of tool-use data, we propose a general tool-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":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2505.16410","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-22T09:00:19Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"ee4c09314112ba9653b44106347c736fae8d3352c15b73774c46e73693a9473a","abstract_canon_sha256":"1f0ccba605213c990a74cd863a47439cd13aad8b6878b67e4c32dd0d9f599959"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:07:35.033628Z","signature_b64":"Sf+xGH2DfTgSVPXjvEsWJ7+xzYwI77RkdrdiNCCXPo/tCSEpKBSwLVWSujifabcAN1suh3YDINgUgwvBBTrWCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7e9a9a8a6485900271c4146faab80f7536f491c52d7415ea2ed5f55285921c9b","last_reissued_at":"2026-07-05T11:07:35.033078Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:07:35.033078Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Tool-Star: Empowering LLM-Brained Multi-Tool Reasoner via Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Guanting Dong, Guorui Zhou, Hangyu Mao, Hongjin Qian, Jiajie Jin, Ji-Rong Wen, Xiaoxi Li, Yifei Chen, Yutao Zhu, Zhicheng Dou","submitted_at":"2025-05-22T09:00:19Z","abstract_excerpt":"Recently, large language models (LLMs) have shown remarkable reasoning capabilities via large-scale reinforcement learning (RL). However, leveraging the RL algorithm to empower effective multi-tool collaborative reasoning in LLMs remains an open challenge. In this paper, we introduce Tool-Star, an RL-based framework designed to empower LLMs to autonomously invoke multiple external tools during stepwise reasoning. Tool-Star integrates six types of tools and incorporates systematic designs in both data synthesis and training. To address the scarcity of tool-use data, we propose a general tool-in"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.16410","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2505.16410/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"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":"2505.16410","created_at":"2026-07-05T11:07:35.033151+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.16410v1","created_at":"2026-07-05T11:07:35.033151+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.16410","created_at":"2026-07-05T11:07:35.033151+00:00"},{"alias_kind":"pith_short_12","alias_value":"P2NJVCTEQWIA","created_at":"2026-07-05T11:07:35.033151+00:00"},{"alias_kind":"pith_short_16","alias_value":"P2NJVCTEQWIAE4OE","created_at":"2026-07-05T11:07:35.033151+00:00"},{"alias_kind":"pith_short_8","alias_value":"P2NJVCTE","created_at":"2026-07-05T11:07:35.033151+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":22,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.12384","citing_title":"APPO: Agentic Procedural Policy Optimization","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05885","citing_title":"When Denser Credit Is Not Enough: Evidence-Calibrated Policy Optimization for Long-Horizon LLM Agent Training","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03762","citing_title":"Tool-Aware Optimization with Entropy Guidance for Efficient Agentic Reinforcement Learning","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02518","citing_title":"ToolFG: Towards Well-Grounded Fine-Grained Image Classification","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01249","citing_title":"Trust Region On-Policy Distillation","ref_index":118,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02132","citing_title":"Learning When Not to Act: Mitigating Tool Abuse in Agentic Reinforcement Learning","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20743","citing_title":"Draw2Think: Harnessing Geometry Reasoning through Constraint Engine Interaction","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2509.08827","citing_title":"A Survey of Reinforcement Learning for Large Reasoning Models","ref_index":113,"is_internal_anchor":false},{"citing_arxiv_id":"2508.07407","citing_title":"A Comprehensive Survey of Self-Evolving AI Agents: A New Paradigm Bridging Foundation Models and Lifelong Agentic Systems","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2603.16876","citing_title":"Multi-Modal Multi-Agent Reinforcement Learning for Radiology Report Generation","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2603.23964","citing_title":"From Pixels to Digital Agents: An Empirical Study on the Taxonomy and Technological Trends of Reinforcement Learning Environments","ref_index":201,"is_internal_anchor":false},{"citing_arxiv_id":"2507.13334","citing_title":"A Survey of Context Engineering for Large Language Models","ref_index":231,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12004","citing_title":"Learning Agentic Policy from Action Guidance","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12481","citing_title":"ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09544","citing_title":"TIDE-Bench: Task-Aware and Diagnostic Evaluation of Tool-Integrated Reasoning","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07237","citing_title":"Teaching Language Models to Think in Code","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09931","citing_title":"PruneTIR: Inference-Time Tool Call Pruning for Effective yet Efficient Tool-Integrated Reasoning","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09813","citing_title":"Controllable and Verifiable Tool-Use Data Synthesis for Agentic Reinforcement Learning","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09455","citing_title":"E3-TIR: Enhanced Experience Exploitation for Tool-Integrated Reasoning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07237","citing_title":"Teaching Language Models to Think in Code","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05387","citing_title":"Data-Driven Function Calling Improvements in Large Language Model for Online Financial QA","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18292","citing_title":"Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligence","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU","json":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU.json","graph_json":"https://pith.science/api/pith-number/P2NJVCTEQWIAE4OECRX2VOAPOU/graph.json","events_json":"https://pith.science/api/pith-number/P2NJVCTEQWIAE4OECRX2VOAPOU/events.json","paper":"https://pith.science/paper/P2NJVCTE"},"agent_actions":{"view_html":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU","download_json":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU.json","view_paper":"https://pith.science/paper/P2NJVCTE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.16410&json=true","fetch_graph":"https://pith.science/api/pith-number/P2NJVCTEQWIAE4OECRX2VOAPOU/graph.json","fetch_events":"https://pith.science/api/pith-number/P2NJVCTEQWIAE4OECRX2VOAPOU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU/action/storage_attestation","attest_author":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU/action/author_attestation","sign_citation":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU/action/citation_signature","submit_replication":"https://pith.science/pith/P2NJVCTEQWIAE4OECRX2VOAPOU/action/replication_record"}},"created_at":"2026-07-05T11:07:35.033151+00:00","updated_at":"2026-07-05T11:07:35.033151+00:00"}