{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4ENXHK473AGORLRXCLGCSMJBHE","short_pith_number":"pith:4ENXHK47","schema_version":"1.0","canonical_sha256":"e11b73ab9fd80ce8ae3712cc293121390d21ca508d693fa65e1e02c678a83bdc","source":{"kind":"arxiv","id":"2401.10415","version":2},"attestation_state":"computed","paper":{"title":"Can Large Language Model Summarizers Adapt to Diverse Scientific Communication Goals?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Marcio Fonseca, Shay B. Cohen","submitted_at":"2024-01-18T23:00:54Z","abstract_excerpt":"In this work, we investigate the controllability of large language models (LLMs) on scientific summarization tasks. We identify key stylistic and content coverage factors that characterize different types of summaries such as paper reviews, abstracts, and lay summaries. By controlling stylistic features, we find that non-fine-tuned LLMs outperform humans in the MuP review generation task, both in terms of similarity to reference summaries and human preferences. Also, we show that we can improve the controllability of LLMs with keyword-based classifier-free guidance (CFG) while achieving lexica"},"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":"2401.10415","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-18T23:00:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0fdb4d57c765113d99ff27677791a4df845e651370d5033bf1a8edfb6dbb9f55","abstract_canon_sha256":"d048cb9cb550ed1edbe81ecfc17cf0cb57fa93d1c7fb77797090f29c8e00596f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:37:16.655063Z","signature_b64":"4YkcrJCxGWOjQQAVakfPOzdf2HysEvjn9XUykjrr5OhUbtQudMGo2YTlXBMtVncaEkjFuqXgiu04SJcXkbkODg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e11b73ab9fd80ce8ae3712cc293121390d21ca508d693fa65e1e02c678a83bdc","last_reissued_at":"2026-07-05T08:37:16.654653Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:37:16.654653Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Can Large Language Model Summarizers Adapt to Diverse Scientific Communication Goals?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Marcio Fonseca, Shay B. Cohen","submitted_at":"2024-01-18T23:00:54Z","abstract_excerpt":"In this work, we investigate the controllability of large language models (LLMs) on scientific summarization tasks. We identify key stylistic and content coverage factors that characterize different types of summaries such as paper reviews, abstracts, and lay summaries. By controlling stylistic features, we find that non-fine-tuned LLMs outperform humans in the MuP review generation task, both in terms of similarity to reference summaries and human preferences. Also, we show that we can improve the controllability of LLMs with keyword-based classifier-free guidance (CFG) while achieving lexica"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.10415","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/2401.10415/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":"2401.10415","created_at":"2026-07-05T08:37:16.654707+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.10415v2","created_at":"2026-07-05T08:37:16.654707+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.10415","created_at":"2026-07-05T08:37:16.654707+00:00"},{"alias_kind":"pith_short_12","alias_value":"4ENXHK473AGO","created_at":"2026-07-05T08:37:16.654707+00:00"},{"alias_kind":"pith_short_16","alias_value":"4ENXHK473AGORLRX","created_at":"2026-07-05T08:37:16.654707+00:00"},{"alias_kind":"pith_short_8","alias_value":"4ENXHK47","created_at":"2026-07-05T08:37:16.654707+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.01265","citing_title":"Beyond In-Context Learning: Aligning Long-form Generation of Large Language Models via Task-Inherent Attribute Guidelines","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE","json":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE.json","graph_json":"https://pith.science/api/pith-number/4ENXHK473AGORLRXCLGCSMJBHE/graph.json","events_json":"https://pith.science/api/pith-number/4ENXHK473AGORLRXCLGCSMJBHE/events.json","paper":"https://pith.science/paper/4ENXHK47"},"agent_actions":{"view_html":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE","download_json":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE.json","view_paper":"https://pith.science/paper/4ENXHK47","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.10415&json=true","fetch_graph":"https://pith.science/api/pith-number/4ENXHK473AGORLRXCLGCSMJBHE/graph.json","fetch_events":"https://pith.science/api/pith-number/4ENXHK473AGORLRXCLGCSMJBHE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE/action/storage_attestation","attest_author":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE/action/author_attestation","sign_citation":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE/action/citation_signature","submit_replication":"https://pith.science/pith/4ENXHK473AGORLRXCLGCSMJBHE/action/replication_record"}},"created_at":"2026-07-05T08:37:16.654707+00:00","updated_at":"2026-07-05T08:37:16.654707+00:00"}