{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:RTWVQODRY3CHQ3PX55EE4AWE7V","short_pith_number":"pith:RTWVQODR","schema_version":"1.0","canonical_sha256":"8ced583871c6c4786df7ef484e02c4fd726eabcf4d59deeaacbf23fb58e082b7","source":{"kind":"arxiv","id":"2603.21438","version":2},"attestation_state":"computed","paper":{"title":"PROMPT2BOX: Uncovering Entailment Structure among LLM Prompts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andrew McCallum, Haw-Shiuan Chang, Neeladri Bhuiya, Shib Sankar Dasgupta","submitted_at":"2026-03-22T22:59:18Z","abstract_excerpt":"To discover the weaknesses of LLMs, researchers often embed prompts into a vector space and cluster them to extract insightful patterns. However, vector embeddings primarily capture topical similarity. As a result, prompts that share a topic but differ in specificity, and consequently in difficulty, are often represented similarly, making fine-grained weakness analysis difficult. To address this limitation, we propose PROMPT2BOX, which embeds prompts into a box embedding space using a trained encoder. The encoder, trained on existing and synthesized datasets, outputs box embeddings that captur"},"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":"2603.21438","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-03-22T22:59:18Z","cross_cats_sorted":[],"title_canon_sha256":"1a0920f8d464885a24501c366cf330cfd51e598358711ad3a3fd068fe4c40712","abstract_canon_sha256":"256688caccec787f0d09666689ea9eaa62e6d5c1be06c78983fb408eee6a6fc5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T03:13:56.561709Z","signature_b64":"ns0gRLiapXk04QbLpDEGA0Wboctwn8Y7Zk6Akakm6xn1GR6NAF2ov19FqiSB9vq87opzBAR38je5cBdwm7UxBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8ced583871c6c4786df7ef484e02c4fd726eabcf4d59deeaacbf23fb58e082b7","last_reissued_at":"2026-06-23T03:13:56.561299Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T03:13:56.561299Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PROMPT2BOX: Uncovering Entailment Structure among LLM Prompts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andrew McCallum, Haw-Shiuan Chang, Neeladri Bhuiya, Shib Sankar Dasgupta","submitted_at":"2026-03-22T22:59:18Z","abstract_excerpt":"To discover the weaknesses of LLMs, researchers often embed prompts into a vector space and cluster them to extract insightful patterns. However, vector embeddings primarily capture topical similarity. As a result, prompts that share a topic but differ in specificity, and consequently in difficulty, are often represented similarly, making fine-grained weakness analysis difficult. To address this limitation, we propose PROMPT2BOX, which embeds prompts into a box embedding space using a trained encoder. The encoder, trained on existing and synthesized datasets, outputs box embeddings that captur"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.21438","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/2603.21438/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":"2603.21438","created_at":"2026-06-23T03:13:56.561366+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.21438v2","created_at":"2026-06-23T03:13:56.561366+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.21438","created_at":"2026-06-23T03:13:56.561366+00:00"},{"alias_kind":"pith_short_12","alias_value":"RTWVQODRY3CH","created_at":"2026-06-23T03:13:56.561366+00:00"},{"alias_kind":"pith_short_16","alias_value":"RTWVQODRY3CHQ3PX","created_at":"2026-06-23T03:13:56.561366+00:00"},{"alias_kind":"pith_short_8","alias_value":"RTWVQODR","created_at":"2026-06-23T03:13:56.561366+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/RTWVQODRY3CHQ3PX55EE4AWE7V","json":"https://pith.science/pith/RTWVQODRY3CHQ3PX55EE4AWE7V.json","graph_json":"https://pith.science/api/pith-number/RTWVQODRY3CHQ3PX55EE4AWE7V/graph.json","events_json":"https://pith.science/api/pith-number/RTWVQODRY3CHQ3PX55EE4AWE7V/events.json","paper":"https://pith.science/paper/RTWVQODR"},"agent_actions":{"view_html":"https://pith.science/pith/RTWVQODRY3CHQ3PX55EE4AWE7V","download_json":"https://pith.science/pith/RTWVQODRY3CHQ3PX55EE4AWE7V.json","view_paper":"https://pith.science/paper/RTWVQODR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.21438&json=true","fetch_graph":"https://pith.science/api/pith-number/RTWVQODRY3CHQ3PX55EE4AWE7V/graph.json","fetch_events":"https://pith.science/api/pith-number/RTWVQODRY3CHQ3PX55EE4AWE7V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RTWVQODRY3CHQ3PX55EE4AWE7V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RTWVQODRY3CHQ3PX55EE4AWE7V/action/storage_attestation","attest_author":"https://pith.science/pith/RTWVQODRY3CHQ3PX55EE4AWE7V/action/author_attestation","sign_citation":"https://pith.science/pith/RTWVQODRY3CHQ3PX55EE4AWE7V/action/citation_signature","submit_replication":"https://pith.science/pith/RTWVQODRY3CHQ3PX55EE4AWE7V/action/replication_record"}},"created_at":"2026-06-23T03:13:56.561366+00:00","updated_at":"2026-06-23T03:13:56.561366+00:00"}