{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KYQOPQBG42HNBOQOVSUI2WLNSM","short_pith_number":"pith:KYQOPQBG","schema_version":"1.0","canonical_sha256":"5620e7c026e68ed0ba0eaca88d596d9339b4eada0ed869b0b924724d74a6627d","source":{"kind":"arxiv","id":"2403.11085","version":4},"attestation_state":"computed","paper":{"title":"m&m's: A Benchmark to Evaluate Tool-Use for multi-step multi-modal Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Jieyu Zhang, Ranjay Krishna, Tanmay Gupta, Weikai Huang, Zixian Ma","submitted_at":"2024-03-17T04:36:18Z","abstract_excerpt":"Real-world multi-modal problems are rarely solved by a single machine learning model, and often require multi-step computational plans that involve stitching several models. Tool-augmented LLMs hold tremendous promise for automating the generation of such computational plans. However, the lack of standardized benchmarks for evaluating LLMs as planners for multi-step multi-modal tasks has prevented a systematic study of planner design decisions. Should LLMs generate a full plan in a single shot or step-by-step? Should they invoke tools directly with Python code or through structured data format"},"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":"2403.11085","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-03-17T04:36:18Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"ee56b7e58e340285a682d8cae6488d8ef0ac533f9d1acad0953bbf5c71a8fcd2","abstract_canon_sha256":"2f91b318dd6a334fc52f008051ddd86522ba91c25762a92ac4a7f4f509d23713"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:09:57.431673Z","signature_b64":"y25WpdbHgu7zm8fCBVCGhlJ2XmEbTi/+bfFjJc7kIZZIn756kxhYbIOT9v3j4LUNB0gL3WQ5v0bFK+4wb3p5BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5620e7c026e68ed0ba0eaca88d596d9339b4eada0ed869b0b924724d74a6627d","last_reissued_at":"2026-07-05T09:09:57.431189Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:09:57.431189Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"m&m's: A Benchmark to Evaluate Tool-Use for multi-step multi-modal Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Jieyu Zhang, Ranjay Krishna, Tanmay Gupta, Weikai Huang, Zixian Ma","submitted_at":"2024-03-17T04:36:18Z","abstract_excerpt":"Real-world multi-modal problems are rarely solved by a single machine learning model, and often require multi-step computational plans that involve stitching several models. Tool-augmented LLMs hold tremendous promise for automating the generation of such computational plans. However, the lack of standardized benchmarks for evaluating LLMs as planners for multi-step multi-modal tasks has prevented a systematic study of planner design decisions. Should LLMs generate a full plan in a single shot or step-by-step? Should they invoke tools directly with Python code or through structured data format"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.11085","kind":"arxiv","version":4},"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/2403.11085/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":"2403.11085","created_at":"2026-07-05T09:09:57.431242+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.11085v4","created_at":"2026-07-05T09:09:57.431242+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.11085","created_at":"2026-07-05T09:09:57.431242+00:00"},{"alias_kind":"pith_short_12","alias_value":"KYQOPQBG42HN","created_at":"2026-07-05T09:09:57.431242+00:00"},{"alias_kind":"pith_short_16","alias_value":"KYQOPQBG42HNBOQO","created_at":"2026-07-05T09:09:57.431242+00:00"},{"alias_kind":"pith_short_8","alias_value":"KYQOPQBG","created_at":"2026-07-05T09:09:57.431242+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.17012","citing_title":"Sustainability assessment using multimodal AI agents","ref_index":35,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM","json":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM.json","graph_json":"https://pith.science/api/pith-number/KYQOPQBG42HNBOQOVSUI2WLNSM/graph.json","events_json":"https://pith.science/api/pith-number/KYQOPQBG42HNBOQOVSUI2WLNSM/events.json","paper":"https://pith.science/paper/KYQOPQBG"},"agent_actions":{"view_html":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM","download_json":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM.json","view_paper":"https://pith.science/paper/KYQOPQBG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.11085&json=true","fetch_graph":"https://pith.science/api/pith-number/KYQOPQBG42HNBOQOVSUI2WLNSM/graph.json","fetch_events":"https://pith.science/api/pith-number/KYQOPQBG42HNBOQOVSUI2WLNSM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM/action/storage_attestation","attest_author":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM/action/author_attestation","sign_citation":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM/action/citation_signature","submit_replication":"https://pith.science/pith/KYQOPQBG42HNBOQOVSUI2WLNSM/action/replication_record"}},"created_at":"2026-07-05T09:09:57.431242+00:00","updated_at":"2026-07-05T09:09:57.431242+00:00"}