{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:H3G3ESSLNYJ7URUZPAF3DGYIEQ","short_pith_number":"pith:H3G3ESSL","schema_version":"1.0","canonical_sha256":"3ecdb24a4b6e13fa4699780bb19b082417a0f480045e2aad80d3e3156b2c7364","source":{"kind":"arxiv","id":"2605.26661","version":1},"attestation_state":"computed","paper":{"title":"Respecting Modality Gap in Post-hoc Out-of-distribution Detection with Pre-trained Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bo Peng, Jie Lu, Ling Chen, Yadan Luo, Yuanwei Hu, Zhen Fang","submitted_at":"2026-05-26T07:46:23Z","abstract_excerpt":"Out-of-distribution (OOD) detection has emerged as a popular technique to enhance the reliability of machine learning models by identifying unexpected inputs from unknown classes. Recent progress in pre-trained vision-language models (VLMs) has enabled zero-shot OOD detection without access to in-distribution (ID) training data; in this setting, existing methods commonly treat text embeddings of class names as class prototypes. In this paper, we challenge the widely adopted text-as-prototype paradigm by theoretically showing that off-the-shelf textual prototypes are generally misaligned with t"},"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":"2605.26661","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-26T07:46:23Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b1ce2abbf00fe8583009f6f259124408c4e76d09e596a7e57399790ef316b6b5","abstract_canon_sha256":"4daf41b813e8428d07a04f5745cec4f77a6fd8fd707962a280b9d35517c7f0aa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:06:04.567466Z","signature_b64":"+BR4huB6KWrQUc8m5TZLQMThxnWAKm8zCItDPOrwpDETQ5M4eQGh2M+kY4Su8ELN9EuGJwj8UtRLXUKaRKFbCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3ecdb24a4b6e13fa4699780bb19b082417a0f480045e2aad80d3e3156b2c7364","last_reissued_at":"2026-05-27T01:06:04.566738Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:06:04.566738Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Respecting Modality Gap in Post-hoc Out-of-distribution Detection with Pre-trained Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bo Peng, Jie Lu, Ling Chen, Yadan Luo, Yuanwei Hu, Zhen Fang","submitted_at":"2026-05-26T07:46:23Z","abstract_excerpt":"Out-of-distribution (OOD) detection has emerged as a popular technique to enhance the reliability of machine learning models by identifying unexpected inputs from unknown classes. Recent progress in pre-trained vision-language models (VLMs) has enabled zero-shot OOD detection without access to in-distribution (ID) training data; in this setting, existing methods commonly treat text embeddings of class names as class prototypes. In this paper, we challenge the widely adopted text-as-prototype paradigm by theoretically showing that off-the-shelf textual prototypes are generally misaligned with t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.26661","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/2605.26661/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":"2605.26661","created_at":"2026-05-27T01:06:04.566854+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.26661v1","created_at":"2026-05-27T01:06:04.566854+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.26661","created_at":"2026-05-27T01:06:04.566854+00:00"},{"alias_kind":"pith_short_12","alias_value":"H3G3ESSLNYJ7","created_at":"2026-05-27T01:06:04.566854+00:00"},{"alias_kind":"pith_short_16","alias_value":"H3G3ESSLNYJ7URUZ","created_at":"2026-05-27T01:06:04.566854+00:00"},{"alias_kind":"pith_short_8","alias_value":"H3G3ESSL","created_at":"2026-05-27T01:06:04.566854+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/H3G3ESSLNYJ7URUZPAF3DGYIEQ","json":"https://pith.science/pith/H3G3ESSLNYJ7URUZPAF3DGYIEQ.json","graph_json":"https://pith.science/api/pith-number/H3G3ESSLNYJ7URUZPAF3DGYIEQ/graph.json","events_json":"https://pith.science/api/pith-number/H3G3ESSLNYJ7URUZPAF3DGYIEQ/events.json","paper":"https://pith.science/paper/H3G3ESSL"},"agent_actions":{"view_html":"https://pith.science/pith/H3G3ESSLNYJ7URUZPAF3DGYIEQ","download_json":"https://pith.science/pith/H3G3ESSLNYJ7URUZPAF3DGYIEQ.json","view_paper":"https://pith.science/paper/H3G3ESSL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.26661&json=true","fetch_graph":"https://pith.science/api/pith-number/H3G3ESSLNYJ7URUZPAF3DGYIEQ/graph.json","fetch_events":"https://pith.science/api/pith-number/H3G3ESSLNYJ7URUZPAF3DGYIEQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/H3G3ESSLNYJ7URUZPAF3DGYIEQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/H3G3ESSLNYJ7URUZPAF3DGYIEQ/action/storage_attestation","attest_author":"https://pith.science/pith/H3G3ESSLNYJ7URUZPAF3DGYIEQ/action/author_attestation","sign_citation":"https://pith.science/pith/H3G3ESSLNYJ7URUZPAF3DGYIEQ/action/citation_signature","submit_replication":"https://pith.science/pith/H3G3ESSLNYJ7URUZPAF3DGYIEQ/action/replication_record"}},"created_at":"2026-05-27T01:06:04.566854+00:00","updated_at":"2026-05-27T01:06:04.566854+00:00"}