{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:E6KPDYDQPBKKXTJXQMEX34RPIK","short_pith_number":"pith:E6KPDYDQ","schema_version":"1.0","canonical_sha256":"2794f1e0707854abcd3783097df22f42beeab7d8be40dce882367a80611813b4","source":{"kind":"arxiv","id":"2411.16679","version":2},"attestation_state":"computed","paper":{"title":"Do Large Language Models Perform Latent Multi-Hop Reasoning without Exploiting Shortcuts?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Elena Gribovskaya, Mor Geva, Nora Kassner, Sebastian Riedel, Sohee Yang","submitted_at":"2024-11-25T18:59:30Z","abstract_excerpt":"We evaluate how well Large Language Models (LLMs) latently recall and compose facts to answer multi-hop queries like \"In the year Scarlett Johansson was born, the Summer Olympics were hosted in the country of\". One major challenge in such evaluation is that LLMs may have developed shortcuts by encountering the head entity \"Scarlett Johansson\" and the answer entity \"United States\" in the same training sequences or merely guess the answer based on frequency-based priors. To prevent shortcuts, we exclude test queries where the head and answer entities might have co-appeared during training. Throu"},"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":"2411.16679","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-25T18:59:30Z","cross_cats_sorted":[],"title_canon_sha256":"6d443f0f7a4c78d45aad0552ed2514296de148bfede8a4d6ddeaaf870560e76e","abstract_canon_sha256":"77c3a7b4b9214428ea8b6cb6d72ec8905f2e56a638277bfa136d17b098b074d8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:13:15.937312Z","signature_b64":"tMR0/jg38JWDRIXxYe39k930ruWJWXZgsziMe7KV6RUToEU892ZtONwT+oR4PLq6Tm7vCVvfYx4V0/mxQhf7AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2794f1e0707854abcd3783097df22f42beeab7d8be40dce882367a80611813b4","last_reissued_at":"2026-07-05T11:13:15.936763Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:13:15.936763Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do Large Language Models Perform Latent Multi-Hop Reasoning without Exploiting Shortcuts?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Elena Gribovskaya, Mor Geva, Nora Kassner, Sebastian Riedel, Sohee Yang","submitted_at":"2024-11-25T18:59:30Z","abstract_excerpt":"We evaluate how well Large Language Models (LLMs) latently recall and compose facts to answer multi-hop queries like \"In the year Scarlett Johansson was born, the Summer Olympics were hosted in the country of\". One major challenge in such evaluation is that LLMs may have developed shortcuts by encountering the head entity \"Scarlett Johansson\" and the answer entity \"United States\" in the same training sequences or merely guess the answer based on frequency-based priors. To prevent shortcuts, we exclude test queries where the head and answer entities might have co-appeared during training. Throu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.16679","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/2411.16679/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":"2411.16679","created_at":"2026-07-05T11:13:15.936831+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.16679v2","created_at":"2026-07-05T11:13:15.936831+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.16679","created_at":"2026-07-05T11:13:15.936831+00:00"},{"alias_kind":"pith_short_12","alias_value":"E6KPDYDQPBKK","created_at":"2026-07-05T11:13:15.936831+00:00"},{"alias_kind":"pith_short_16","alias_value":"E6KPDYDQPBKKXTJX","created_at":"2026-07-05T11:13:15.936831+00:00"},{"alias_kind":"pith_short_8","alias_value":"E6KPDYDQ","created_at":"2026-07-05T11:13:15.936831+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07157","citing_title":"Think Fast: Estimating No-CoT Task-Completion Time Horizons of Frontier AI Models","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07157","citing_title":"Think Fast: Estimating No-CoT Task-Completion Time Horizons of Frontier AI Models","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK","json":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK.json","graph_json":"https://pith.science/api/pith-number/E6KPDYDQPBKKXTJXQMEX34RPIK/graph.json","events_json":"https://pith.science/api/pith-number/E6KPDYDQPBKKXTJXQMEX34RPIK/events.json","paper":"https://pith.science/paper/E6KPDYDQ"},"agent_actions":{"view_html":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK","download_json":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK.json","view_paper":"https://pith.science/paper/E6KPDYDQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.16679&json=true","fetch_graph":"https://pith.science/api/pith-number/E6KPDYDQPBKKXTJXQMEX34RPIK/graph.json","fetch_events":"https://pith.science/api/pith-number/E6KPDYDQPBKKXTJXQMEX34RPIK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK/action/storage_attestation","attest_author":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK/action/author_attestation","sign_citation":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK/action/citation_signature","submit_replication":"https://pith.science/pith/E6KPDYDQPBKKXTJXQMEX34RPIK/action/replication_record"}},"created_at":"2026-07-05T11:13:15.936831+00:00","updated_at":"2026-07-05T11:13:15.936831+00:00"}