{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ZFO7MPWZC5R4XHO42SB6CTELEG","short_pith_number":"pith:ZFO7MPWZ","schema_version":"1.0","canonical_sha256":"c95df63ed91763cb9ddcd483e14c8b21bccd2c045fc3d843f5cd412950ad0f34","source":{"kind":"arxiv","id":"2408.04682","version":2},"attestation_state":"computed","paper":{"title":"ToolSandbox: A Stateful, Conversational, Interactive Evaluation Benchmark for LLM Tool Use Capabilities","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bernhard Aumayer, Felix Bai, Feng Nan, Guoli Yin, Jiarui Lu, Mengyu Li, Ruoming Pang, Shen Ma, Shuang Ma, Thomas Holleis, Yizhe Zhang, Zirui Wang","submitted_at":"2024-08-08T05:45:42Z","abstract_excerpt":"Recent large language models (LLMs) advancements sparked a growing research interest in tool assisted LLMs solving real-world challenges, which calls for comprehensive evaluation of tool-use capabilities. While previous works focused on either evaluating over stateless web services (RESTful API), based on a single turn user prompt, or an off-policy dialog trajectory, ToolSandbox includes stateful tool execution, implicit state dependencies between tools, a built-in user simulator supporting on-policy conversational evaluation and a dynamic evaluation strategy for intermediate and final milesto"},"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":"2408.04682","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-08-08T05:45:42Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"2a180a412fe973b53acb8764d943e156def70045fe688f8463619779bb803b3f","abstract_canon_sha256":"26a9d323522a5b20e3a351cbd42733f525779c1ab375660c4968a7363a0a360e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:50:08.946223Z","signature_b64":"MJv7zWPs6qDrlIonQEL4OLaZp/uYIq7DvXG21zN1mgGMpriN48zrW6ZQ8DPMGB6HQJjpUZjrOX2l4ZjKJsL/Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c95df63ed91763cb9ddcd483e14c8b21bccd2c045fc3d843f5cd412950ad0f34","last_reissued_at":"2026-07-05T10:50:08.945725Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:50:08.945725Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ToolSandbox: A Stateful, Conversational, Interactive Evaluation Benchmark for LLM Tool Use Capabilities","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bernhard Aumayer, Felix Bai, Feng Nan, Guoli Yin, Jiarui Lu, Mengyu Li, Ruoming Pang, Shen Ma, Shuang Ma, Thomas Holleis, Yizhe Zhang, Zirui Wang","submitted_at":"2024-08-08T05:45:42Z","abstract_excerpt":"Recent large language models (LLMs) advancements sparked a growing research interest in tool assisted LLMs solving real-world challenges, which calls for comprehensive evaluation of tool-use capabilities. While previous works focused on either evaluating over stateless web services (RESTful API), based on a single turn user prompt, or an off-policy dialog trajectory, ToolSandbox includes stateful tool execution, implicit state dependencies between tools, a built-in user simulator supporting on-policy conversational evaluation and a dynamic evaluation strategy for intermediate and final milesto"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.04682","kind":"arxiv","version":2},"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/2408.04682/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":"2408.04682","created_at":"2026-07-05T10:50:08.945785+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.04682v2","created_at":"2026-07-05T10:50:08.945785+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.04682","created_at":"2026-07-05T10:50:08.945785+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZFO7MPWZC5R4","created_at":"2026-07-05T10:50:08.945785+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZFO7MPWZC5R4XHO4","created_at":"2026-07-05T10:50:08.945785+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZFO7MPWZ","created_at":"2026-07-05T10:50:08.945785+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18051","citing_title":"Compositional Skill Routing for LLM Agents: Decompose, Retrieve, and Compose","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01465","citing_title":"Beyond Next-Token Prediction: An RLVR Proof of Concept for Tool-Use Agents on Atlassian Workflows","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02965","citing_title":"Designing for Doubt: The Case for Informed Abstention in Autonomous Agents","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27141","citing_title":"VitaBench 2.0: Evaluating Personalized and Proactive Agents in Long-Term User Interactions","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22643","citing_title":"Boiling the Frog: A Multi-Turn Benchmark for Agentic Safety","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2603.14987","citing_title":"Beyond Benchmark Islands: Toward Representative Trustworthiness Evaluation for Agentic AI","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22643","citing_title":"Boiling the Frog: A Multi-Turn Benchmark for Agentic Safety","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2510.07043","citing_title":"COMPASS: Benchmarking Constrained Optimization in LLM Agents","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2507.21046","citing_title":"A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence","ref_index":90,"is_internal_anchor":false},{"citing_arxiv_id":"2410.09024","citing_title":"AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2503.23278","citing_title":"Model Context Protocol (MCP): Landscape, Security Threats, and Future Research Directions","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25318","citing_title":"Cutscene Agent: An LLM Agent Framework for Automated 3D Cutscene Generation","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2506.07982","citing_title":"$\\tau^2$-Bench: Evaluating Conversational Agents in a Dual-Control Environment","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12116","citing_title":"The A-R Behavioral Space: Execution-Level Profiling of Tool-Using Language Model Agents in Organizational Deployment","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG","json":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG.json","graph_json":"https://pith.science/api/pith-number/ZFO7MPWZC5R4XHO42SB6CTELEG/graph.json","events_json":"https://pith.science/api/pith-number/ZFO7MPWZC5R4XHO42SB6CTELEG/events.json","paper":"https://pith.science/paper/ZFO7MPWZ"},"agent_actions":{"view_html":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG","download_json":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG.json","view_paper":"https://pith.science/paper/ZFO7MPWZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.04682&json=true","fetch_graph":"https://pith.science/api/pith-number/ZFO7MPWZC5R4XHO42SB6CTELEG/graph.json","fetch_events":"https://pith.science/api/pith-number/ZFO7MPWZC5R4XHO42SB6CTELEG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG/action/storage_attestation","attest_author":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG/action/author_attestation","sign_citation":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG/action/citation_signature","submit_replication":"https://pith.science/pith/ZFO7MPWZC5R4XHO42SB6CTELEG/action/replication_record"}},"created_at":"2026-07-05T10:50:08.945785+00:00","updated_at":"2026-07-05T10:50:08.945785+00:00"}