{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "yao-et-al-tau-bench-2024", "title": "Yao and colleagues: tau-bench tool-agent-user reliability evaluation", "summary": "Evaluated policy-following API agents in conversations with simulated users, checking final database states and repeated-trial reliability.", "boundary": "The user is model-simulated and domains are constrained. Repeated-trial pass^k differs from best-of-k pass@k; final-state checks do not prove every interaction is safe or appropriate.", "area": "AI evaluation and agent infrastructure", "topic": "Artificial intelligence", "year": 2024, "lane": "HUMAN", "result": "BENCHMARK", "contribution_kind": "EVALUATION_RESOURCE_DESIGN", "package_id": "yao-et-al-tau-bench-2024", "target": { "record_id": "urn:uuid:f6fe9d66-3d22-4fe4-a2db-ac264f9e6add", "version": 1, "content_hash": "sha256:8e09a34c2209cf41ef506dae300bfd946415e0698433b668cf86a16300fd8d7c" }, "humans": [ "Shunyu Yao", "Noah Shinn", "Pedram Razavi", "Karthik Narasimhan" ], "authors_complete": true, "human_role_note": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.", "agents": [], "related": [], "subresults": [], "scientific_assessment": { "source_review": "SOURCE_CURATED", "scientific_review": "NOT_REVIEWED_BY_WRS", "formal_replay": "NOT_RUN", "provider_attestation": "NONE" }, "sources": [ { "id": "cur20261015-arxiv-2406-12045v1", "title": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "url": "https://arxiv.org/abs/2406.12045v1" } ], "points": "65000", "published_grade": { "id": "yao-et-al-tau-bench-2024", "entry_id": "yao-et-al-tau-bench-2024", "band": "B", "modifier": "+", "target": { "record_id": "urn:uuid:f6fe9d66-3d22-4fe4-a2db-ac264f9e6add", "version": 1, "content_hash": "sha256:8e09a34c2209cf41ef506dae300bfd946415e0698433b668cf86a16300fd8d7c" }, "what_changed": "Evaluated policy-following API agents in conversations with simulated users, checking final database states and repeated-trial reliability.", "scope": "Evaluated policy-following API agents in conversations with simulated users, checking final database states and repeated-trial reliability. The user is model-simulated and domains are constrained. Repeated-trial pass^k differs from best-of-k pass@k; final-state checks do not prove every interaction is safe or appropriate.", "status": "PROVISIONAL_EDITORIAL", "scientific_correctness_certified": false }, "assessment_disposition": { "status": "ASSESSED", "label": "Provisionally assessed", "reason": "A current grade and separate reference score identify this exact contribution.", "target": { "record_id": "urn:uuid:f6fe9d66-3d22-4fe4-a2db-ac264f9e6add", "version": 1, "content_hash": "sha256:8e09a34c2209cf41ef506dae300bfd946415e0698433b668cf86a16300fd8d7c" }, "edition": "ratings-e039106f30dec177" } }, "version": { "kind": "PUBLIC_RECORD_VIEW", "previous": null, "published_at": "2026-10-16T00:00:00Z", "publisher": { "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33", "role": "WRS_EDITORIAL_PUBLICATION" }, "record": { "contributors": [ { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Shunyu Yao", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "url": "https://arxiv.org/abs/2406.12045v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Noah Shinn", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "url": "https://arxiv.org/abs/2406.12045v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Pedram Razavi", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "url": "https://arxiv.org/abs/2406.12045v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Karthik Narasimhan", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "url": "https://arxiv.org/abs/2406.12045v1" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "url": "https://arxiv.org/abs/2406.12045v1" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "The user is model-simulated and domains are constrained. Repeated-trial pass^k differs from best-of-k pass@k; final-state checks do not prove every interaction is safe or appropriate.", "domain": "AI evaluation and agent infrastructure" }, "sources": [ { "citation": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "url": "https://arxiv.org/abs/2406.12045v1" } ], "statement": "# Yao and colleagues: tau-bench tool-agent-user reliability evaluation\n\nEvaluated policy-following API agents in conversations with simulated users, checking final database states and repeated-trial reliability.\n\n## Scientific scope\n\nThe user is model-simulated and domains are constrained. Repeated-trial pass^k differs from best-of-k pass@k; final-state checks do not prove every interaction is safe or appropriate.\n\n## Reading status\n\nThis is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.\n\n## People and systems\n\nShunyu Yao; Noah Shinn; Pedram Razavi; Karthik Narasimhan\n\nSource-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.\n\n\n\n## Original sources\n\n- [$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains](https://arxiv.org/abs/2406.12045v1)\n", "summary": "Evaluated policy-following API agents in conversations with simulated users, checking final database states and repeated-trial reliability.", "title": "Yao and colleagues: tau-bench tool-agent-user reliability evaluation" }, "record_id": "urn:uuid:f6fe9d66-3d22-4fe4-a2db-ac264f9e6add", "version": 1, "source_ref": { "record_id": "urn:uuid:f6fe9d66-3d22-4fe4-a2db-ac264f9e6add", "version": 1, "content_hash": "sha256:8e09a34c2209cf41ef506dae300bfd946415e0698433b668cf86a16300fd8d7c" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:4ab0d5aa49e4593bdd10fb626d8488f4627d1b33b2e1c0ac50a5735c58b2fe4b" }, "target": { "record_id": "urn:uuid:f6fe9d66-3d22-4fe4-a2db-ac264f9e6add", "version": 1, "content_hash": "sha256:8e09a34c2209cf41ef506dae300bfd946415e0698433b668cf86a16300fd8d7c" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "yao-et-al-tau-bench-2024", "entry_id": "yao-et-al-tau-bench-2024", "revision": 1, "title": "Yao and colleagues: tau-bench tool-agent-user reliability evaluation", "context": "AI evaluation and agent infrastructure", "track": "Evaluation resource design", "band": "B", "modifier": "+", "status": "PROVISIONAL_EDITORIAL", "assessed_on": "2026-10-16", "evidence_cutoff": "2026-10-16", "achievement_state": "SUPPORTED_SCOPED_RESULT", "what_changed": "Evaluated policy-following API agents in conversations with simulated users, checking final database states and repeated-trial reliability.", "scope": "Evaluated policy-following API agents in conversations with simulated users, checking final database states and repeated-trial reliability. The user is model-simulated and domains are constrained. Repeated-trial pass^k differs from best-of-k pass@k; final-state checks do not prove every interaction is safe or appropriate.", "evidence_summary": "The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. The user is model-simulated and domains are constrained. Repeated-trial pass^k differs from best-of-k pass@k; final-state checks do not prove every interaction is safe or appropriate.", "references": [ { "title": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "url": "https://arxiv.org/abs/2406.12045v1", "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification." } ], "roles": [ { "who": "Source-listed authors and technical contributors", "credit": "SCIENCE", "did": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities." } ], "target": { "record_id": "urn:uuid:f6fe9d66-3d22-4fe4-a2db-ac264f9e6add", "version": 1, "content_hash": "sha256:8e09a34c2209cf41ef506dae300bfd946415e0698433b668cf86a16300fd8d7c" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "65000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": null, "reading_status": "SOURCE_LINKED_NOTE" }