{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BCR6YZTAORDN4IDBFM25JX3VFU","short_pith_number":"pith:BCR6YZTA","schema_version":"1.0","canonical_sha256":"08a3ec66607446de20612b35d4df752d220d0af371aae34b130a6384943f6d63","source":{"kind":"arxiv","id":"2506.23719","version":1},"attestation_state":"computed","paper":{"title":"DABstep: Data Agent Benchmark for Multi-step Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Alex Egg, Andreu Mora, Friso Kingma, Leandro Von Werra, Martin Iglesias Goyanes, Thomas Wolf","submitted_at":"2025-06-30T10:49:21Z","abstract_excerpt":"We introduce DABstep, a novel benchmark for evaluating AI agents on realistic multi-step data analysis tasks. DABstep comprises over 450 real-world challenges derived from a financial analytics platform, requiring models to combine code-based data processing with contextual reasoning over heterogeneous documentation. Each task demands an iterative, multi-step problem-solving approach, testing capabilities in data manipulation, cross-referencing multiple sources, and precise result reporting. The benchmark provides a factoid-style answer format with automatic correctness checks for objective sc"},"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":"2506.23719","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-30T10:49:21Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1b423777e86a71908dc7e60e402261597efca1d0dd84990eb55761fac22626b2","abstract_canon_sha256":"b8cf4cc3b49e174b55c1e23c07db764ad7bdc4b37539e413b3ee4d5b08f8a287"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:29:27.018393Z","signature_b64":"NIVH49JGtWaaBPXgrH0r+ZbfiBzGPor5RMUo4jqoYf/zSB9VE8+5jCx5ObHrFs7x2YwPsH+M8srMzmM0L/EcDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"08a3ec66607446de20612b35d4df752d220d0af371aae34b130a6384943f6d63","last_reissued_at":"2026-07-05T11:29:27.017831Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:29:27.017831Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DABstep: Data Agent Benchmark for Multi-step Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Alex Egg, Andreu Mora, Friso Kingma, Leandro Von Werra, Martin Iglesias Goyanes, Thomas Wolf","submitted_at":"2025-06-30T10:49:21Z","abstract_excerpt":"We introduce DABstep, a novel benchmark for evaluating AI agents on realistic multi-step data analysis tasks. DABstep comprises over 450 real-world challenges derived from a financial analytics platform, requiring models to combine code-based data processing with contextual reasoning over heterogeneous documentation. Each task demands an iterative, multi-step problem-solving approach, testing capabilities in data manipulation, cross-referencing multiple sources, and precise result reporting. The benchmark provides a factoid-style answer format with automatic correctness checks for objective sc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.23719","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/2506.23719/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":"2506.23719","created_at":"2026-07-05T11:29:27.017896+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.23719v1","created_at":"2026-07-05T11:29:27.017896+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.23719","created_at":"2026-07-05T11:29:27.017896+00:00"},{"alias_kind":"pith_short_12","alias_value":"BCR6YZTAORDN","created_at":"2026-07-05T11:29:27.017896+00:00"},{"alias_kind":"pith_short_16","alias_value":"BCR6YZTAORDN4IDB","created_at":"2026-07-05T11:29:27.017896+00:00"},{"alias_kind":"pith_short_8","alias_value":"BCR6YZTA","created_at":"2026-07-05T11:29:27.017896+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":13,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.16000","citing_title":"GRACE-DS: a Guarded Reward-guided Agent Correction Environment in Data Science","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00384","citing_title":"VESTA: Visual Exploration with Statistical Tool Agents","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02503","citing_title":"DataClawBench: An Agent Benchmark for Exploratory Real-World Financial Data Analysis","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26297","citing_title":"Agentic AI Workload Characteristics","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30434","citing_title":"LongDS-Bench: On the Failure of Long-Horizon Agentic Data Analysis","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06416","citing_title":"Unsupervised Skill Discovery for Agentic Data Analysis","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21338","citing_title":"Text Analytics Evaluation Framework: A Case Study on LLMs and Social Media","ref_index":77,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02503","citing_title":"DataClawBench: An Agent Benchmark for Exploratory Real-World Financial Data Analysis","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2512.13168","citing_title":"Finch: Benchmarking Finance & Accounting across Spreadsheet-Centric Enterprise Workflows","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08687","citing_title":"PrepBench: How Far Are We from Natural-Language-Driven Data Preparation?","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10516","citing_title":"Structure-Grounded Knowledge Retrieval via Code Dependencies for Multi-Step Data Reasoning","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16682","citing_title":"KAIROS: Stateful, Context-Aware Power-Efficient Agentic Inference Serving","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02503","citing_title":"DataClawBench: An Agent Benchmark for Exploratory Real-World Financial Data Analysis","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU","json":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU.json","graph_json":"https://pith.science/api/pith-number/BCR6YZTAORDN4IDBFM25JX3VFU/graph.json","events_json":"https://pith.science/api/pith-number/BCR6YZTAORDN4IDBFM25JX3VFU/events.json","paper":"https://pith.science/paper/BCR6YZTA"},"agent_actions":{"view_html":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU","download_json":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU.json","view_paper":"https://pith.science/paper/BCR6YZTA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.23719&json=true","fetch_graph":"https://pith.science/api/pith-number/BCR6YZTAORDN4IDBFM25JX3VFU/graph.json","fetch_events":"https://pith.science/api/pith-number/BCR6YZTAORDN4IDBFM25JX3VFU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU/action/storage_attestation","attest_author":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU/action/author_attestation","sign_citation":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU/action/citation_signature","submit_replication":"https://pith.science/pith/BCR6YZTAORDN4IDBFM25JX3VFU/action/replication_record"}},"created_at":"2026-07-05T11:29:27.017896+00:00","updated_at":"2026-07-05T11:29:27.017896+00:00"}