{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KYGMLU46GQNDYOZU52LLRQKM5B","short_pith_number":"pith:KYGMLU46","schema_version":"1.0","canonical_sha256":"560cc5d39e341a3c3b34ee96b8c14ce85e02429ccfc6e747aa30b30b2d6b246c","source":{"kind":"arxiv","id":"2401.01989","version":3},"attestation_state":"computed","paper":{"title":"Revisiting Zero-Shot Abstractive Summarization in the Era of Large Language Models from the Perspective of Position Bias","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Anshuman Chhabra, Hadi Askari, Prasant Mohapatra","submitted_at":"2024-01-03T21:38:40Z","abstract_excerpt":"We characterize and study zero-shot abstractive summarization in Large Language Models (LLMs) by measuring position bias, which we propose as a general formulation of the more restrictive lead bias phenomenon studied previously in the literature. Position bias captures the tendency of a model unfairly prioritizing information from certain parts of the input text over others, leading to undesirable behavior. Through numerous experiments on four diverse real-world datasets, we study position bias in multiple LLM models such as GPT 3.5-Turbo, Llama-2, and Dolly-v2, as well as state-of-the-art pre"},"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":"2401.01989","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-03T21:38:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"51b4e1d87c99183633e71f8f45b70461ba4fd69b9cc45d9fc94251a34ef89a0a","abstract_canon_sha256":"600cc197792e0fd765b667e6ca3b35befd41d9ed728a98508484f08b987cadab"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:57:48.362606Z","signature_b64":"x7EJfPo2xpu/+GDRsxG2LdXwlC3Mo4BRzJtEeH8Wngx2RcrCdpDA+Z1cpb3OsqBj2iVGii2sf/oNsG+WOphDDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"560cc5d39e341a3c3b34ee96b8c14ce85e02429ccfc6e747aa30b30b2d6b246c","last_reissued_at":"2026-07-05T07:57:48.362124Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:57:48.362124Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Revisiting Zero-Shot Abstractive Summarization in the Era of Large Language Models from the Perspective of Position Bias","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Anshuman Chhabra, Hadi Askari, Prasant Mohapatra","submitted_at":"2024-01-03T21:38:40Z","abstract_excerpt":"We characterize and study zero-shot abstractive summarization in Large Language Models (LLMs) by measuring position bias, which we propose as a general formulation of the more restrictive lead bias phenomenon studied previously in the literature. Position bias captures the tendency of a model unfairly prioritizing information from certain parts of the input text over others, leading to undesirable behavior. Through numerous experiments on four diverse real-world datasets, we study position bias in multiple LLM models such as GPT 3.5-Turbo, Llama-2, and Dolly-v2, as well as state-of-the-art pre"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.01989","kind":"arxiv","version":3},"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/2401.01989/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":"2401.01989","created_at":"2026-07-05T07:57:48.362182+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.01989v3","created_at":"2026-07-05T07:57:48.362182+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.01989","created_at":"2026-07-05T07:57:48.362182+00:00"},{"alias_kind":"pith_short_12","alias_value":"KYGMLU46GQND","created_at":"2026-07-05T07:57:48.362182+00:00"},{"alias_kind":"pith_short_16","alias_value":"KYGMLU46GQNDYOZU","created_at":"2026-07-05T07:57:48.362182+00:00"},{"alias_kind":"pith_short_8","alias_value":"KYGMLU46","created_at":"2026-07-05T07:57:48.362182+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.08662","citing_title":"RoToR: Towards More Reliable Responses for Order-Invariant Inputs","ref_index":9,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B","json":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B.json","graph_json":"https://pith.science/api/pith-number/KYGMLU46GQNDYOZU52LLRQKM5B/graph.json","events_json":"https://pith.science/api/pith-number/KYGMLU46GQNDYOZU52LLRQKM5B/events.json","paper":"https://pith.science/paper/KYGMLU46"},"agent_actions":{"view_html":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B","download_json":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B.json","view_paper":"https://pith.science/paper/KYGMLU46","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.01989&json=true","fetch_graph":"https://pith.science/api/pith-number/KYGMLU46GQNDYOZU52LLRQKM5B/graph.json","fetch_events":"https://pith.science/api/pith-number/KYGMLU46GQNDYOZU52LLRQKM5B/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B/action/storage_attestation","attest_author":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B/action/author_attestation","sign_citation":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B/action/citation_signature","submit_replication":"https://pith.science/pith/KYGMLU46GQNDYOZU52LLRQKM5B/action/replication_record"}},"created_at":"2026-07-05T07:57:48.362182+00:00","updated_at":"2026-07-05T07:57:48.362182+00:00"}