{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:M5IDYKG7R5ERZ7BWQBXPUUHETF","short_pith_number":"pith:M5IDYKG7","schema_version":"1.0","canonical_sha256":"67503c28df8f491cfc36806efa50e4994880f5f57bd4c49a7680c772547d3d91","source":{"kind":"arxiv","id":"2509.04512","version":1},"attestation_state":"computed","paper":{"title":"Scaling behavior of large language models in emotional safety classification across sizes and tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Andr\\'e Ferreira Castro, Edoardo Pinzuti, Oliver T\\\"uscher","submitted_at":"2025-09-02T20:53:03Z","abstract_excerpt":"Understanding how large language models (LLMs) process emotionally sensitive content is critical for building safe and reliable systems, particularly in mental health contexts. We investigate the scaling behavior of LLMs on two key tasks: trinary classification of emotional safety (safe vs. unsafe vs. borderline) and multi-label classification using a six-category safety risk taxonomy. To support this, we construct a novel dataset by merging several human-authored mental health datasets (> 15K samples) and augmenting them with emotion re-interpretation prompts generated via ChatGPT. We evaluat"},"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":"2509.04512","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-02T20:53:03Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"24020023a2e57c0650e7f1bd717681f49a9fbb028046c1a78cc1de6b4341048e","abstract_canon_sha256":"fc12f9bb24a31f2bc0353f9feaca65955959e59f603927ca91b1dbed2b6c8ec9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:05:24.422393Z","signature_b64":"6IFM4A3eHsuT75r4zoOx9ONQzy0L3IR177hn1aoVG2SFByiFR1sU3zOSJFGWZ6ULegn5PykQvscbKh249AxSBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"67503c28df8f491cfc36806efa50e4994880f5f57bd4c49a7680c772547d3d91","last_reissued_at":"2026-07-05T12:05:24.421833Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:05:24.421833Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Scaling behavior of large language models in emotional safety classification across sizes and tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Andr\\'e Ferreira Castro, Edoardo Pinzuti, Oliver T\\\"uscher","submitted_at":"2025-09-02T20:53:03Z","abstract_excerpt":"Understanding how large language models (LLMs) process emotionally sensitive content is critical for building safe and reliable systems, particularly in mental health contexts. We investigate the scaling behavior of LLMs on two key tasks: trinary classification of emotional safety (safe vs. unsafe vs. borderline) and multi-label classification using a six-category safety risk taxonomy. To support this, we construct a novel dataset by merging several human-authored mental health datasets (> 15K samples) and augmenting them with emotion re-interpretation prompts generated via ChatGPT. We evaluat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.04512","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/2509.04512/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":"2509.04512","created_at":"2026-07-05T12:05:24.421894+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.04512v1","created_at":"2026-07-05T12:05:24.421894+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.04512","created_at":"2026-07-05T12:05:24.421894+00:00"},{"alias_kind":"pith_short_12","alias_value":"M5IDYKG7R5ER","created_at":"2026-07-05T12:05:24.421894+00:00"},{"alias_kind":"pith_short_16","alias_value":"M5IDYKG7R5ERZ7BW","created_at":"2026-07-05T12:05:24.421894+00:00"},{"alias_kind":"pith_short_8","alias_value":"M5IDYKG7","created_at":"2026-07-05T12:05:24.421894+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/M5IDYKG7R5ERZ7BWQBXPUUHETF","json":"https://pith.science/pith/M5IDYKG7R5ERZ7BWQBXPUUHETF.json","graph_json":"https://pith.science/api/pith-number/M5IDYKG7R5ERZ7BWQBXPUUHETF/graph.json","events_json":"https://pith.science/api/pith-number/M5IDYKG7R5ERZ7BWQBXPUUHETF/events.json","paper":"https://pith.science/paper/M5IDYKG7"},"agent_actions":{"view_html":"https://pith.science/pith/M5IDYKG7R5ERZ7BWQBXPUUHETF","download_json":"https://pith.science/pith/M5IDYKG7R5ERZ7BWQBXPUUHETF.json","view_paper":"https://pith.science/paper/M5IDYKG7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.04512&json=true","fetch_graph":"https://pith.science/api/pith-number/M5IDYKG7R5ERZ7BWQBXPUUHETF/graph.json","fetch_events":"https://pith.science/api/pith-number/M5IDYKG7R5ERZ7BWQBXPUUHETF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M5IDYKG7R5ERZ7BWQBXPUUHETF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M5IDYKG7R5ERZ7BWQBXPUUHETF/action/storage_attestation","attest_author":"https://pith.science/pith/M5IDYKG7R5ERZ7BWQBXPUUHETF/action/author_attestation","sign_citation":"https://pith.science/pith/M5IDYKG7R5ERZ7BWQBXPUUHETF/action/citation_signature","submit_replication":"https://pith.science/pith/M5IDYKG7R5ERZ7BWQBXPUUHETF/action/replication_record"}},"created_at":"2026-07-05T12:05:24.421894+00:00","updated_at":"2026-07-05T12:05:24.421894+00:00"}