{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:HLB5V2EMI6GL3WMLZYGGYVCSRT","short_pith_number":"pith:HLB5V2EM","canonical_record":{"source":{"id":"2605.13450","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T12:46:05Z","cross_cats_sorted":["cs.CL","cs.HC"],"title_canon_sha256":"f89e3dc824ca299a0e7bf4fa92d54273981e405c905e82b1f2891db6a56888cd","abstract_canon_sha256":"efe89762f839231939802c28a6b3aa2575fe6709ea712e4e09e6301d4e378b4b"},"schema_version":"1.0"},"canonical_sha256":"3ac3dae88c478cbdd98bce0c6c54528cc037fe7b12cd0811d6bfad2ace532773","source":{"kind":"arxiv","id":"2605.13450","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13450","created_at":"2026-05-18T02:44:41Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13450v1","created_at":"2026-05-18T02:44:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13450","created_at":"2026-05-18T02:44:41Z"},{"alias_kind":"pith_short_12","alias_value":"HLB5V2EMI6GL","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"HLB5V2EMI6GL3WML","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"HLB5V2EM","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:HLB5V2EMI6GL3WMLZYGGYVCSRT","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13450","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T12:46:05Z","cross_cats_sorted":["cs.CL","cs.HC"],"title_canon_sha256":"f89e3dc824ca299a0e7bf4fa92d54273981e405c905e82b1f2891db6a56888cd","abstract_canon_sha256":"efe89762f839231939802c28a6b3aa2575fe6709ea712e4e09e6301d4e378b4b"},"schema_version":"1.0"},"canonical_sha256":"3ac3dae88c478cbdd98bce0c6c54528cc037fe7b12cd0811d6bfad2ace532773","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:41.907034Z","signature_b64":"GJzn+wasOzd6ikieUHAx3bH3u7F+K8OPG2taP7I1c2kBJfMeJOhSyOZg4o1thqRH26y8CNdUfTOGEyuADbrwAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3ac3dae88c478cbdd98bce0c6c54528cc037fe7b12cd0811d6bfad2ace532773","last_reissued_at":"2026-05-18T02:44:41.906529Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:41.906529Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13450","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"i404xbp3F4jljupMstydAj0kyzcwPwPrngwaU5IYMsZWSYGigfoqtI1DT9WIs5pWREFPC3Hey3Nx1vo+PSdnBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T06:35:31.690528Z"},"content_sha256":"1934cc6b3ac3a54e922edf88403d2951732f917186763977d3f4617d9276403e","schema_version":"1.0","event_id":"sha256:1934cc6b3ac3a54e922edf88403d2951732f917186763977d3f4617d9276403e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:HLB5V2EMI6GL3WMLZYGGYVCSRT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Assessing the Creativity of Large Language Models: Testing, Limits, and New Frontiers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"The Divergent Remote Association Test predicts large language models' scientific ideation ability where other creativity tests fail.","cross_cats":["cs.CL","cs.HC"],"primary_cat":"cs.AI","authors_text":"Alexi Gladstone, Heng Ji, Jonah Black, Samuel Schapiro","submitted_at":"2026-05-13T12:46:05Z","abstract_excerpt":"Measuring the creativity of large language models (LLMs) is essential for designing methods that can improve creativity and for enhancing our scientific understanding of this ability. To accomplish this, it has become common in recent years to administer tests of human creativity to LLMs. Although these tests provide a convenient and fully automated way to score \"creativity,\" their validity as measures of machine creativity has not been established, and these tests already have limited validity as predictors of human creativity. To address this problem, we conduct the first large-scale, system"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The DRAT is the first and only creativity test for LLMs that is a significant predictor of scientific ideation ability, demonstrating robustness across major design choices. Furthermore, the performance gain of the DRAT is not recoverable from any linear combination of the Divergent Association Task and the Remote Associates Test.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the human creativity tests and the new DRAT validly measure the target constructs of creative writing, divergent thinking, and scientific ideation in LLMs, despite the abstract noting that these tests have limited validity even for humans.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"The Divergent Remote Association Test (DRAT) is the first creativity test that significantly predicts LLMs' scientific ideation ability, unlike prior tests such as DAT or RAT.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"The Divergent Remote Association Test predicts large language models' scientific ideation ability where other creativity tests fail.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4db5302998971fb09bc730272be3e20e6353de882a9d591fa1246a7116c19888"},"source":{"id":"2605.13450","kind":"arxiv","version":1},"verdict":{"id":"515c9a1e-956f-40c4-8216-0657137edbdc","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:06:08.475895Z","strongest_claim":"The DRAT is the first and only creativity test for LLMs that is a significant predictor of scientific ideation ability, demonstrating robustness across major design choices. Furthermore, the performance gain of the DRAT is not recoverable from any linear combination of the Divergent Association Task and the Remote Associates Test.","one_line_summary":"The Divergent Remote Association Test (DRAT) is the first creativity test that significantly predicts LLMs' scientific ideation ability, unlike prior tests such as DAT or RAT.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the human creativity tests and the new DRAT validly measure the target constructs of creative writing, divergent thinking, and scientific ideation in LLMs, despite the abstract noting that these tests have limited validity even for humans.","pith_extraction_headline":"The Divergent Remote Association Test predicts large language models' scientific ideation ability where other creativity tests fail."},"references":{"count":29,"sample":[{"doi":"10.1016/j.tsc.2021.100859","year":2021,"title":"doi: 10.1016/j.tsc.2021.100859","work_id":"66f8a07c-4069-447f-8b30-9b886fa2d2c2","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"20 Chiang, W.-L., Zheng, L., Sheng, Y., Angelopoulos, A","work_id":"76d45713-09ee-4c29-9b8c-de9c43e6b182","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1911,"title":"On the Measure of Intelligence","work_id":"d8980a59-aa48-447b-8852-b7aca2b41b2c","ref_index":3,"cited_arxiv_id":"1911.01547","is_internal_anchor":true},{"doi":"10.3758/s13423-018-1517-7","year":null,"title":"doi: 10.3758/s13423-018-1517-7","work_id":"57fb9235-09f9-4fea-a399-30c7f96e3247","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"CresOWLve: Benchmarking Creative Problem-Solving Over Real-World Knowledge","work_id":"0205e64a-0b55-4afa-9d36-0c802d940969","ref_index":5,"cited_arxiv_id":"2604.03374","is_internal_anchor":true}],"resolved_work":29,"snapshot_sha256":"a8a20ea337fb8e0db352493f435cc6919a92a498fe61c427a8d2a77ec8e984de","internal_anchors":5},"formal_canon":{"evidence_count":1,"snapshot_sha256":"c42a42c716dda6ea6591d54d4da85604f0dd9658eb90edb70507edf33e4e8e1a"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"515c9a1e-956f-40c4-8216-0657137edbdc"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6mlliJMGs10mLRIvWI3QazzBu79Bfl0F0J4GWMxHehds7xp2QV12L7MkGzpBqZWE3A3nSkxk4XOgiRhphC/WDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T06:35:31.691058Z"},"content_sha256":"7c13f3af65d190797b40f462c175cc9d3620b950563573ae821ed4d2eabd0150","schema_version":"1.0","event_id":"sha256:7c13f3af65d190797b40f462c175cc9d3620b950563573ae821ed4d2eabd0150"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:HLB5V2EMI6GL3WMLZYGGYVCSRT","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.18653/v1/2024.acl-long.18.urlhttp://arxiv.org/abs/2305.14259http) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"Wang, Q., Downey, D., Ji, H., and Hope, T. SciMON: Scientific Inspiration Machines Optimized for Novelty.62024a.doi: 10.18653/v1/2024.acl-long.18.URL http://arxiv.org/abs/2305.14259http: //dx.doi.org/10.18653/v1/2024.acl-long.18. Wang, Y., ","arxiv_id":"2605.13450","detector":"doi_compliance","evidence":{"ref_index":23,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"Wang, Q., Downey, D., Ji, H., and Hope, T. SciMON: Scientific Inspiration Machines Optimized for Novelty.62024a.doi: 10.18653/v1/2024.acl-long.18.URL http://arxiv.org/abs/2305.14259http: //dx.doi.org/10.18653/v1/2024.acl-long.18. Wang, Y., ","reconstructed_doi":"10.18653/v1/2024.acl-long.18.urlhttp://arxiv.org/abs/2305.14259http"},"severity":"advisory","ref_index":23,"audited_at":"2026-05-19T06:32:40.065429Z","event_type":"pith.integrity.v1","detected_doi":"10.18653/v1/2024.acl-long.18.urlhttp://arxiv.org/abs/2305.14259http","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"da8e426fa99261561745899435940efdef7c676cf6ce8bc1d597238f1967f793","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.0.0","detected_arxiv_id":null,"integrity_event_id":163,"payload_sha256":"f766e5e9ad798b2ecdd7a620482ea7dbdc7db88cf4df4d45a841ac341e4d4929","signature_b64":"YSirvlkahBmBmI171JBmaceYdLW+H69EMh/5ah+iaaghp21jsGI0ABGjWInvI7W1874jgkr31qCx8S3/SeAXBQ==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-19T06:36:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"zJxPRTs6qfWoY2arLSI2OU4AChOiE38+R/YFZKXM04vtQE4AN0qmuHkEyY68wjuuXTUxEh5DjvWBKZBTZ5OrBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T06:35:31.691947Z"},"content_sha256":"e1edad4de62422ead27ac1fa3695d291979fded4b072261daecbde0f47c599fb","schema_version":"1.0","event_id":"sha256:e1edad4de62422ead27ac1fa3695d291979fded4b072261daecbde0f47c599fb"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:HLB5V2EMI6GL3WMLZYGGYVCSRT","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1073/pnas.2022340118/-/DCSupplemental.y.Olsson) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"doi: 10.1073/pnas.2022340118/-/DCSupplemental.y. Olsson, C., Elhage, N., Nanda, N., Joseph, N., DasSarma, N., Henighan, T., Mann, B., Askell, A., Bai, Y., Chen, A., Conerly, T., Drain, D., Ganguli, D., Hatfield-Dodds, Z., Hernandez, D., Joh","arxiv_id":"2605.13450","detector":"doi_compliance","evidence":{"ref_index":12,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"doi: 10.1073/pnas.2022340118/-/DCSupplemental.y. Olsson, C., Elhage, N., Nanda, N., Joseph, N., DasSarma, N., Henighan, T., Mann, B., Askell, A., Bai, Y., Chen, A., Conerly, T., Drain, D., Ganguli, D., Hatfield-Dodds, Z., Hernandez, D., Joh","reconstructed_doi":"10.1073/pnas.2022340118/-/DCSupplemental.y.Olsson"},"severity":"advisory","ref_index":12,"audited_at":"2026-05-19T06:32:40.065429Z","event_type":"pith.integrity.v1","detected_doi":"10.1073/pnas.2022340118/-/DCSupplemental.y.Olsson","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"d0c35f265771c168fccf71a0bc3e1425780450a84ae0c9da2dd1ec3dde100677","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.0.0","detected_arxiv_id":null,"integrity_event_id":162,"payload_sha256":"ad1145f97d36f7de345611852cb73dedd5e62b4029c2d6162d64dced0d8f0f72","signature_b64":"gXilAxcuAV+jTrvlKHw0EBDNi7frIolMA95+9ZYQbJeLPYwBJ8otebDCprZIEliJ7p7KaB9yE4CzZ/svD0S5AQ==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-19T06:36:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SlZfgD1HrFVsXesmd2Xwf4fKFSZ//7136Ro50dL+4XAvF2pqKnpkErygZHYbTuBp5X4f74b99IffaKRPHJdrBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T06:35:31.692231Z"},"content_sha256":"dc0513339a4a31af8053f72eefb355216bab864325ce511b81ef2ac84dc4f81c","schema_version":"1.0","event_id":"sha256:dc0513339a4a31af8053f72eefb355216bab864325ce511b81ef2ac84dc4f81c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HLB5V2EMI6GL3WMLZYGGYVCSRT/bundle.json","state_url":"https://pith.science/pith/HLB5V2EMI6GL3WMLZYGGYVCSRT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HLB5V2EMI6GL3WMLZYGGYVCSRT/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-21T06:35:31Z","links":{"resolver":"https://pith.science/pith/HLB5V2EMI6GL3WMLZYGGYVCSRT","bundle":"https://pith.science/pith/HLB5V2EMI6GL3WMLZYGGYVCSRT/bundle.json","state":"https://pith.science/pith/HLB5V2EMI6GL3WMLZYGGYVCSRT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HLB5V2EMI6GL3WMLZYGGYVCSRT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:HLB5V2EMI6GL3WMLZYGGYVCSRT","merge_version":"pith-open-graph-merge-v1","event_count":4,"valid_event_count":4,"invalid_event_count":0,"equivocation_count":1,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"efe89762f839231939802c28a6b3aa2575fe6709ea712e4e09e6301d4e378b4b","cross_cats_sorted":["cs.CL","cs.HC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T12:46:05Z","title_canon_sha256":"f89e3dc824ca299a0e7bf4fa92d54273981e405c905e82b1f2891db6a56888cd"},"schema_version":"1.0","source":{"id":"2605.13450","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13450","created_at":"2026-05-18T02:44:41Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13450v1","created_at":"2026-05-18T02:44:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13450","created_at":"2026-05-18T02:44:41Z"},{"alias_kind":"pith_short_12","alias_value":"HLB5V2EMI6GL","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"HLB5V2EMI6GL3WML","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"HLB5V2EM","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:7c13f3af65d190797b40f462c175cc9d3620b950563573ae821ed4d2eabd0150","target":"graph","created_at":"2026-05-18T02:44:41Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"The DRAT is the first and only creativity test for LLMs that is a significant predictor of scientific ideation ability, demonstrating robustness across major design choices. Furthermore, the performance gain of the DRAT is not recoverable from any linear combination of the Divergent Association Task and the Remote Associates Test."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the human creativity tests and the new DRAT validly measure the target constructs of creative writing, divergent thinking, and scientific ideation in LLMs, despite the abstract noting that these tests have limited validity even for humans."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"The Divergent Remote Association Test (DRAT) is the first creativity test that significantly predicts LLMs' scientific ideation ability, unlike prior tests such as DAT or RAT."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"The Divergent Remote Association Test predicts large language models' scientific ideation ability where other creativity tests fail."}],"snapshot_sha256":"4db5302998971fb09bc730272be3e20e6353de882a9d591fa1246a7116c19888"},"formal_canon":{"evidence_count":1,"snapshot_sha256":"c42a42c716dda6ea6591d54d4da85604f0dd9658eb90edb70507edf33e4e8e1a"},"paper":{"abstract_excerpt":"Measuring the creativity of large language models (LLMs) is essential for designing methods that can improve creativity and for enhancing our scientific understanding of this ability. To accomplish this, it has become common in recent years to administer tests of human creativity to LLMs. Although these tests provide a convenient and fully automated way to score \"creativity,\" their validity as measures of machine creativity has not been established, and these tests already have limited validity as predictors of human creativity. To address this problem, we conduct the first large-scale, system","authors_text":"Alexi Gladstone, Heng Ji, Jonah Black, Samuel Schapiro","cross_cats":["cs.CL","cs.HC"],"headline":"The Divergent Remote Association Test predicts large language models' scientific ideation ability where other creativity tests fail.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T12:46:05Z","title":"Assessing the Creativity of Large Language Models: Testing, Limits, and New Frontiers"},"references":{"count":29,"internal_anchors":5,"resolved_work":29,"sample":[{"cited_arxiv_id":"","doi":"10.1016/j.tsc.2021.100859","is_internal_anchor":false,"ref_index":1,"title":"doi: 10.1016/j.tsc.2021.100859","work_id":"66f8a07c-4069-447f-8b30-9b886fa2d2c2","year":2021},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"20 Chiang, W.-L., Zheng, L., Sheng, Y., Angelopoulos, A","work_id":"76d45713-09ee-4c29-9b8c-de9c43e6b182","year":null},{"cited_arxiv_id":"1911.01547","doi":"","is_internal_anchor":true,"ref_index":3,"title":"On the Measure of Intelligence","work_id":"d8980a59-aa48-447b-8852-b7aca2b41b2c","year":1911},{"cited_arxiv_id":"","doi":"10.3758/s13423-018-1517-7","is_internal_anchor":false,"ref_index":4,"title":"doi: 10.3758/s13423-018-1517-7","work_id":"57fb9235-09f9-4fea-a399-30c7f96e3247","year":null},{"cited_arxiv_id":"2604.03374","doi":"","is_internal_anchor":true,"ref_index":5,"title":"CresOWLve: Benchmarking Creative Problem-Solving Over Real-World Knowledge","work_id":"0205e64a-0b55-4afa-9d36-0c802d940969","year":null}],"snapshot_sha256":"a8a20ea337fb8e0db352493f435cc6919a92a498fe61c427a8d2a77ec8e984de"},"source":{"id":"2605.13450","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:06:08.475895Z","id":"515c9a1e-956f-40c4-8216-0657137edbdc","model_set":{"reader":"grok-4.3"},"one_line_summary":"The Divergent Remote Association Test (DRAT) is the first creativity test that significantly predicts LLMs' scientific ideation ability, unlike prior tests such as DAT or RAT.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"The Divergent Remote Association Test predicts large language models' scientific ideation ability where other creativity tests fail.","strongest_claim":"The DRAT is the first and only creativity test for LLMs that is a significant predictor of scientific ideation ability, demonstrating robustness across major design choices. Furthermore, the performance gain of the DRAT is not recoverable from any linear combination of the Divergent Association Task and the Remote Associates Test.","weakest_assumption":"That the human creativity tests and the new DRAT validly measure the target constructs of creative writing, divergent thinking, and scientific ideation in LLMs, despite the abstract noting that these tests have limited validity even for humans."}},"verdict_id":"515c9a1e-956f-40c4-8216-0657137edbdc"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1934cc6b3ac3a54e922edf88403d2951732f917186763977d3f4617d9276403e","target":"record","created_at":"2026-05-18T02:44:41Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"efe89762f839231939802c28a6b3aa2575fe6709ea712e4e09e6301d4e378b4b","cross_cats_sorted":["cs.CL","cs.HC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T12:46:05Z","title_canon_sha256":"f89e3dc824ca299a0e7bf4fa92d54273981e405c905e82b1f2891db6a56888cd"},"schema_version":"1.0","source":{"id":"2605.13450","kind":"arxiv","version":1}},"canonical_sha256":"3ac3dae88c478cbdd98bce0c6c54528cc037fe7b12cd0811d6bfad2ace532773","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3ac3dae88c478cbdd98bce0c6c54528cc037fe7b12cd0811d6bfad2ace532773","first_computed_at":"2026-05-18T02:44:41.906529Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:41.906529Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"GJzn+wasOzd6ikieUHAx3bH3u7F+K8OPG2taP7I1c2kBJfMeJOhSyOZg4o1thqRH26y8CNdUfTOGEyuADbrwAg==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:41.907034Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13450","source_kind":"arxiv","source_version":1}}},"equivocations":[{"signer_id":"pith.science","event_type":"integrity_finding","target":"integrity","event_ids":["sha256:dc0513339a4a31af8053f72eefb355216bab864325ce511b81ef2ac84dc4f81c","sha256:e1edad4de62422ead27ac1fa3695d291979fded4b072261daecbde0f47c599fb"]}],"invalid_events":[],"applied_event_ids":["sha256:1934cc6b3ac3a54e922edf88403d2951732f917186763977d3f4617d9276403e","sha256:7c13f3af65d190797b40f462c175cc9d3620b950563573ae821ed4d2eabd0150"],"state_sha256":"a5def76f7d991ba341e5750363952a1c8a9d89290a40639a8ace7bd09de46f1c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kjjtGOkNaBAavS+yev1y8Ga+K0L0dJUnBJs4fTWhh68N3TafCXdSNVXMKhaQsaTFgMxg4lBYefdcJ7wk3tH6BA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-21T06:35:31.694979Z","bundle_sha256":"8a611aab9ef6b42d28826572db6ef76ed2e23e241390342a28005106da42f8e7"}}