{"record": {"aliases": ["$\\tau$-Rec: A Verifiable Benchmark for Agentic Recommender Systems"], "artifacts": [{"id": "arxiv:2606.10156", "kind": "paper", "url": "https://arxiv.org/abs/2606.10156"}, {"id": "gh:nbharaths/tau-rec", "kind": "repo", "url": "https://github.com/nbharaths/tau-rec"}], "authors": ["Bharath Sivaram Narasimhan", "Karthik R Narasimhan"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.10156", "source_url": "https://arxiv.org/abs/2606.10156"}, "categories": ["General AI", "Language & Knowledge", "cs.IR"], "description": {"en": "A benchmark for multi-turn agentic recommender systems that exposes dialogue constraints through reveal-tagged elicitation and verifies outcomes against structured catalog predicates."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.10156", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.10156", "title": "τ-Rec"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Princeton University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.10156", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.10156", "source_url": "https://openalex.org/W7164154021"}, "key": "claire-radar:2606.10156", "modality": null, "name": "τ-Rec", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_tau-rec-a-verifiable-benchmark-for-agentic_13bfa014", "origin_source": "arxiv", "recognition_confidence": "0.7", "record_sha256": "a42cd5e36afbfc361c544a992983b51e840b52157a1d380d733954faec0d3cc4", "relation": "introduces", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T17:10:41.268638Z", "source_url": "https://arxiv.org/abs/2606.10156"}, "publisher": null, "released": "2026-06-08", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.10156", "source_url": "https://arxiv.org/abs/2606.10156"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-10156", "source": "claire_radar", "source_benchmark_id": "2606.10156", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["$\\tau$-Rec: A Verifiable Benchmark for Agentic Recommender Systems"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Verifiable agent evaluation, public artifacts, and the reported reliability gap across several model families make it likely to attract near-term interest.", "score": 82}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T17:10:41.268638Z", "inputHash": "fb5e2070dea62608bbd8b5b30484ce2a9a370b3ee9f800c99e31550d2042d620", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "$\\tau$-Rec: A Verifiable Benchmark for Agentic Recommender Systems", "canonicalNameSource": "paper_title", "decisionReason": "The source formally names the benchmark, specifies verifiable evaluation and a pass^k metric, and provides a public code-and-data reuse path.", "evaluationEvidence": {"reuse": "All code and data are publicly available at https://github.com/nbharaths/tau-rec.", "scoring": "employing a pass^k reliability metric", "task": "a benchmark for agentic recommender systems"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T17:10:41.268638Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "A benchmark for multi-turn agentic recommender systems that exposes dialogue constraints through reveal-tagged elicitation and verifies outcomes against structured catalog predicates.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We present $\\tau$-Rec, a benchmark for agentic recommender systems that replaces subjective evaluation with verifiable rewards and a reveal-tagged elicitation (RTE) mechanism that controls how task constraints surface during dialogue."}, "familyId": "bmf_cf0cae3bcbff", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_tau-rec-a-verifiable-benchmark-for-agentic_13bfa014", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/nbharaths/tau-rec", "data": null, "paper": "https://arxiv.org/abs/2606.10156", "pdf": "https://arxiv.org/pdf/2606.10156", "project": null, "report": "https://arxiv.org/abs/2606.10156"}, "metrics": [], "motivation": "As recommender systems transition toward agentic, multi-turn conversational interfaces, evaluation paradigms have struggled to keep pace.", "name": "τ-Rec", "oneLine": "A benchmark for multi-turn agentic recommender systems that exposes dialogue constraints through reveal-tagged elicitation and verifies outcomes against structured catalog predicates.", "paperTitle": "$\\tau$-Rec: A Verifiable Benchmark for Agentic Recommender Systems", "primaryDomain": "General AI", "publication": {"evidence": "Accepted at ACM RecSys 2026 (Reproducibility and Resource Track)", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2606.10156", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "ACM RecSys 2026 (Reproducibility and Resource Track)", "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "publishers": [{"name": "nbharaths", "organizationType": "community", "role": "benchmark-publisher", "sourceUrl": "https://github.com/nbharaths/tau-rec"}], "readiness": "Runnable", "recognitionConfidence": 0.7, "relation": "introduces", "releasedAt": "2026-06-08", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Bharath Sivaram Narasimhan", "Karthik R Narasimhan"], "categories": ["cs.IR", "cs.AI", "cs.CL"], "id": "2606.10156", "title": "$\\tau$-Rec: A Verifiable Benchmark for Agentic Recommender Systems", "type": "arxiv", "url": "https://arxiv.org/abs/2606.10156"}, "sourceUpdatedAt": "2026-07-24T21:13:14Z", "topics": ["cs.IR"], "venueAttempts": [{"decisionRaw": "Accepted at ACM RecSys 2026 (Reproducibility and Resource Track)", "evidence": [{"level": "author-claim", "observedAt": "2026-08-19T11:05:13.395391Z", "rawValue": "Accepted at ACM RecSys 2026 (Reproducibility and Resource Track)", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2606.10156"}], "reviewStatus": "accepted", "venueName": "ACM RecSys 2026 (Reproducibility and Resource Track)"}], "whyItMatters": "It replaces subjective judge-based assessment with verifiable rewards and measures whether recommendation agents reason consistently across repeated trials."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
