{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:APZA4SS2JONR46CJU3DJQSWYYS","short_pith_number":"pith:APZA4SS2","schema_version":"1.0","canonical_sha256":"03f20e4a5a4b9b1e7849a6c6984ad8c48bf62f10a87c6a32e87d47b3e452e5cf","source":{"kind":"arxiv","id":"2502.01676","version":1},"attestation_state":"computed","paper":{"title":"Benchmark on Peer Review Toxic Detection: A Challenging Task with a New Dataset","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CY"],"primary_cat":"cs.CL","authors_text":"Ariadne Dimarogona, Bradley Peterson, Hari Ramalingame, Imon Banerjee, Man Luo, Navya Gangrade, Phillip Howard, Rafael Gan","submitted_at":"2025-02-01T23:01:39Z","abstract_excerpt":"Peer review is crucial for advancing and improving science through constructive criticism. However, toxic feedback can discourage authors and hinder scientific progress. This work explores an important but underexplored area: detecting toxicity in peer reviews. We first define toxicity in peer reviews across four distinct categories and curate a dataset of peer reviews from the OpenReview platform, annotated by human experts according to these definitions. Leveraging this dataset, we benchmark a variety of models, including a dedicated toxicity detection model, a sentiment analysis model, seve"},"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":"2502.01676","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-01T23:01:39Z","cross_cats_sorted":["cs.CY"],"title_canon_sha256":"86ac850381ce92410aa4ac51ed7edbee2bdc0f4df07ff27cb27c4d26ba05b379","abstract_canon_sha256":"13f46892b85060dfd008528fe2ce329fd50635fb071c7ad5edd5df2083f62d2d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:09:17.512982Z","signature_b64":"J6zOObK2g8IZmAxyrQGoT5VAmvmwRP97J5t96q1UhTNHQLgbQ68tNkm1tFbikBeKV5XJvgusEBoLbp1NsRfADA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"03f20e4a5a4b9b1e7849a6c6984ad8c48bf62f10a87c6a32e87d47b3e452e5cf","last_reissued_at":"2026-07-05T10:09:17.512455Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:09:17.512455Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Benchmark on Peer Review Toxic Detection: A Challenging Task with a New Dataset","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CY"],"primary_cat":"cs.CL","authors_text":"Ariadne Dimarogona, Bradley Peterson, Hari Ramalingame, Imon Banerjee, Man Luo, Navya Gangrade, Phillip Howard, Rafael Gan","submitted_at":"2025-02-01T23:01:39Z","abstract_excerpt":"Peer review is crucial for advancing and improving science through constructive criticism. However, toxic feedback can discourage authors and hinder scientific progress. This work explores an important but underexplored area: detecting toxicity in peer reviews. We first define toxicity in peer reviews across four distinct categories and curate a dataset of peer reviews from the OpenReview platform, annotated by human experts according to these definitions. Leveraging this dataset, we benchmark a variety of models, including a dedicated toxicity detection model, a sentiment analysis model, seve"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.01676","kind":"arxiv","version":1},"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/2502.01676/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":"2502.01676","created_at":"2026-07-05T10:09:17.512520+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.01676v1","created_at":"2026-07-05T10:09:17.512520+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.01676","created_at":"2026-07-05T10:09:17.512520+00:00"},{"alias_kind":"pith_short_12","alias_value":"APZA4SS2JONR","created_at":"2026-07-05T10:09:17.512520+00:00"},{"alias_kind":"pith_short_16","alias_value":"APZA4SS2JONR46CJ","created_at":"2026-07-05T10:09:17.512520+00:00"},{"alias_kind":"pith_short_8","alias_value":"APZA4SS2","created_at":"2026-07-05T10:09:17.512520+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS","json":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS.json","graph_json":"https://pith.science/api/pith-number/APZA4SS2JONR46CJU3DJQSWYYS/graph.json","events_json":"https://pith.science/api/pith-number/APZA4SS2JONR46CJU3DJQSWYYS/events.json","paper":"https://pith.science/paper/APZA4SS2"},"agent_actions":{"view_html":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS","download_json":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS.json","view_paper":"https://pith.science/paper/APZA4SS2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.01676&json=true","fetch_graph":"https://pith.science/api/pith-number/APZA4SS2JONR46CJU3DJQSWYYS/graph.json","fetch_events":"https://pith.science/api/pith-number/APZA4SS2JONR46CJU3DJQSWYYS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS/action/storage_attestation","attest_author":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS/action/author_attestation","sign_citation":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS/action/citation_signature","submit_replication":"https://pith.science/pith/APZA4SS2JONR46CJU3DJQSWYYS/action/replication_record"}},"created_at":"2026-07-05T10:09:17.512520+00:00","updated_at":"2026-07-05T10:09:17.512520+00:00"}