{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:3M577TGLWLGCGXZHWV6N2OVO2I","short_pith_number":"pith:3M577TGL","schema_version":"1.0","canonical_sha256":"db3bffcccbb2cc235f27b57cdd3aaed204f80870caf0bb240c445fc7b2027474","source":{"kind":"arxiv","id":"2406.12775","version":2},"attestation_state":"computed","paper":{"title":"Hopping Too Late: Exploring the Limitations of Large Language Models on Multi-Hop Queries","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Amir Globerson, Daniela Gottesman, Eden Biran, Mor Geva, Sohee Yang","submitted_at":"2024-06-18T16:44:13Z","abstract_excerpt":"Large language models (LLMs) can solve complex multi-step problems, but little is known about how these computations are implemented internally. Motivated by this, we study how LLMs answer multi-hop queries such as \"The spouse of the performer of Imagine is\". These queries require two information extraction steps: a latent one for resolving the first hop (\"the performer of Imagine\") into the bridge entity (John Lennon), and another for resolving the second hop (\"the spouse of John Lennon\") into the target entity (Yoko Ono). Understanding how the latent step is computed internally is key to und"},"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":"2406.12775","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-18T16:44:13Z","cross_cats_sorted":[],"title_canon_sha256":"21b783ecf9c19680c19b02197c04186b828150f3bfc0d0a691a33a27da349309","abstract_canon_sha256":"086a4b0f14ee84854a6bda4514eed1dbb5d4beb9f9b0961acf6ce2097fe09249"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:19:59.624040Z","signature_b64":"fVRZqSMPfmGnQZN6G8wbEhLR45y5Ghi6buSE+7RukRBCHYfcvTL51oYcDG93mgWUtm/+4L7sxkfuVwv2nnWKDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"db3bffcccbb2cc235f27b57cdd3aaed204f80870caf0bb240c445fc7b2027474","last_reissued_at":"2026-07-05T09:19:59.623578Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:19:59.623578Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hopping Too Late: Exploring the Limitations of Large Language Models on Multi-Hop Queries","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Amir Globerson, Daniela Gottesman, Eden Biran, Mor Geva, Sohee Yang","submitted_at":"2024-06-18T16:44:13Z","abstract_excerpt":"Large language models (LLMs) can solve complex multi-step problems, but little is known about how these computations are implemented internally. Motivated by this, we study how LLMs answer multi-hop queries such as \"The spouse of the performer of Imagine is\". These queries require two information extraction steps: a latent one for resolving the first hop (\"the performer of Imagine\") into the bridge entity (John Lennon), and another for resolving the second hop (\"the spouse of John Lennon\") into the target entity (Yoko Ono). Understanding how the latent step is computed internally is key to und"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.12775","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/2406.12775/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":"2406.12775","created_at":"2026-07-05T09:19:59.623637+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.12775v2","created_at":"2026-07-05T09:19:59.623637+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.12775","created_at":"2026-07-05T09:19:59.623637+00:00"},{"alias_kind":"pith_short_12","alias_value":"3M577TGLWLGC","created_at":"2026-07-05T09:19:59.623637+00:00"},{"alias_kind":"pith_short_16","alias_value":"3M577TGLWLGCGXZH","created_at":"2026-07-05T09:19:59.623637+00:00"},{"alias_kind":"pith_short_8","alias_value":"3M577TGL","created_at":"2026-07-05T09:19:59.623637+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.13649","citing_title":"Operadic consistency: a label-free signal for compositional reasoning failures in LLMs","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00341","citing_title":"DiscoLoop: Looping Discrete Embeddings and Continuous Hidden States for Multi-hop Reasoning","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2501.19201","citing_title":"Efficient Reasoning with Hidden Thinking","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2510.01685","citing_title":"How Do Language Models Compose Functions?","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08221","citing_title":"NoisyCoconut: Counterfactual Consensus via Latent Space Reasoning","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22951","citing_title":"The Power of Power Law: Asymmetry Enables Compositional Reasoning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2412.06769","citing_title":"Training Large Language Models to Reason in a Continuous Latent Space","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I","json":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I.json","graph_json":"https://pith.science/api/pith-number/3M577TGLWLGCGXZHWV6N2OVO2I/graph.json","events_json":"https://pith.science/api/pith-number/3M577TGLWLGCGXZHWV6N2OVO2I/events.json","paper":"https://pith.science/paper/3M577TGL"},"agent_actions":{"view_html":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I","download_json":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I.json","view_paper":"https://pith.science/paper/3M577TGL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.12775&json=true","fetch_graph":"https://pith.science/api/pith-number/3M577TGLWLGCGXZHWV6N2OVO2I/graph.json","fetch_events":"https://pith.science/api/pith-number/3M577TGLWLGCGXZHWV6N2OVO2I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I/action/storage_attestation","attest_author":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I/action/author_attestation","sign_citation":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I/action/citation_signature","submit_replication":"https://pith.science/pith/3M577TGLWLGCGXZHWV6N2OVO2I/action/replication_record"}},"created_at":"2026-07-05T09:19:59.623637+00:00","updated_at":"2026-07-05T09:19:59.623637+00:00"}