{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:HUB2OSCKYAGWHQ3RNENV6V7DMU","short_pith_number":"pith:HUB2OSCK","schema_version":"1.0","canonical_sha256":"3d03a7484ac00d63c371691b5f57e36535c5624c4af59933606385479b9402f4","source":{"kind":"arxiv","id":"2606.05376","version":1},"attestation_state":"computed","paper":{"title":"SHALA-LLM: Smartly Handling Ambiguous Labels in Aligning LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ashley Wang, Jingyao Wu, Keane Ong, Paul Pu Liang, Rosalind Picard","submitted_at":"2026-06-03T19:25:08Z","abstract_excerpt":"Many human-centered tasks, including natural language inference (NLI) and emotion recognition (ER), have multiple plausible interpretations, leading to label ambiguity and challenging disagreements across human annotators. As LLMs are increasingly deployed in real-world settings, faithfully modeling such ambiguity is essential to identify contested inputs, preserve variability in ambiguous cases, and capture the full distribution of human judgments. Yet, existing LLM alignment approaches have predominantly assumed a single correct label, excluding annotator disagreement during optimization. In"},"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":"2606.05376","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T19:25:08Z","cross_cats_sorted":[],"title_canon_sha256":"a44b3d68cf7f811dbeeed40f775f1f4129fd4c1a709b15560412f24592962808","abstract_canon_sha256":"6cc9bf97a90a5ab67767468e556220bd4b29b9fd29d45aaf4cf96030f04d8384"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T00:13:55.798176Z","signature_b64":"ZwPXSmE1HLieBvtDvJO+Gi6lHcdM4FppE6Zn4zH3SU2jSeHT6uLyE/tKSK1Q50vDm6VNZ08w0aYT2aC30NYNBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3d03a7484ac00d63c371691b5f57e36535c5624c4af59933606385479b9402f4","last_reissued_at":"2026-06-05T00:13:55.797707Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T00:13:55.797707Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SHALA-LLM: Smartly Handling Ambiguous Labels in Aligning LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ashley Wang, Jingyao Wu, Keane Ong, Paul Pu Liang, Rosalind Picard","submitted_at":"2026-06-03T19:25:08Z","abstract_excerpt":"Many human-centered tasks, including natural language inference (NLI) and emotion recognition (ER), have multiple plausible interpretations, leading to label ambiguity and challenging disagreements across human annotators. As LLMs are increasingly deployed in real-world settings, faithfully modeling such ambiguity is essential to identify contested inputs, preserve variability in ambiguous cases, and capture the full distribution of human judgments. Yet, existing LLM alignment approaches have predominantly assumed a single correct label, excluding annotator disagreement during optimization. In"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.05376","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/2606.05376/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":"2606.05376","created_at":"2026-06-05T00:13:55.797775+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.05376v1","created_at":"2026-06-05T00:13:55.797775+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.05376","created_at":"2026-06-05T00:13:55.797775+00:00"},{"alias_kind":"pith_short_12","alias_value":"HUB2OSCKYAGW","created_at":"2026-06-05T00:13:55.797775+00:00"},{"alias_kind":"pith_short_16","alias_value":"HUB2OSCKYAGWHQ3R","created_at":"2026-06-05T00:13:55.797775+00:00"},{"alias_kind":"pith_short_8","alias_value":"HUB2OSCK","created_at":"2026-06-05T00:13:55.797775+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/HUB2OSCKYAGWHQ3RNENV6V7DMU","json":"https://pith.science/pith/HUB2OSCKYAGWHQ3RNENV6V7DMU.json","graph_json":"https://pith.science/api/pith-number/HUB2OSCKYAGWHQ3RNENV6V7DMU/graph.json","events_json":"https://pith.science/api/pith-number/HUB2OSCKYAGWHQ3RNENV6V7DMU/events.json","paper":"https://pith.science/paper/HUB2OSCK"},"agent_actions":{"view_html":"https://pith.science/pith/HUB2OSCKYAGWHQ3RNENV6V7DMU","download_json":"https://pith.science/pith/HUB2OSCKYAGWHQ3RNENV6V7DMU.json","view_paper":"https://pith.science/paper/HUB2OSCK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.05376&json=true","fetch_graph":"https://pith.science/api/pith-number/HUB2OSCKYAGWHQ3RNENV6V7DMU/graph.json","fetch_events":"https://pith.science/api/pith-number/HUB2OSCKYAGWHQ3RNENV6V7DMU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HUB2OSCKYAGWHQ3RNENV6V7DMU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HUB2OSCKYAGWHQ3RNENV6V7DMU/action/storage_attestation","attest_author":"https://pith.science/pith/HUB2OSCKYAGWHQ3RNENV6V7DMU/action/author_attestation","sign_citation":"https://pith.science/pith/HUB2OSCKYAGWHQ3RNENV6V7DMU/action/citation_signature","submit_replication":"https://pith.science/pith/HUB2OSCKYAGWHQ3RNENV6V7DMU/action/replication_record"}},"created_at":"2026-06-05T00:13:55.797775+00:00","updated_at":"2026-06-05T00:13:55.797775+00:00"}