{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:4C7DDSATPBA4DZXMP27FB2D4AH","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"8ec809073586cdb0aab2e3d5d9ac9dfce4883850b688016a82793ea2226a3a01","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2023-05-18T07:50:44Z","title_canon_sha256":"54595f2b8be874df4dfc966c15a4a3810e0c9378c533499c54436055534cba82"},"schema_version":"1.0","source":{"id":"2305.10782","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.10782","created_at":"2026-07-05T07:31:28Z"},{"alias_kind":"arxiv_version","alias_value":"2305.10782v3","created_at":"2026-07-05T07:31:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.10782","created_at":"2026-07-05T07:31:28Z"},{"alias_kind":"pith_short_12","alias_value":"4C7DDSATPBA4","created_at":"2026-07-05T07:31:28Z"},{"alias_kind":"pith_short_16","alias_value":"4C7DDSATPBA4DZXM","created_at":"2026-07-05T07:31:28Z"},{"alias_kind":"pith_short_8","alias_value":"4C7DDSAT","created_at":"2026-07-05T07:31:28Z"}],"graph_snapshots":[{"event_id":"sha256:0405ea7a27e688fd1a3781aca5a493f67d4d70965e47bab10b4d14b1b4bfadf1","target":"graph","created_at":"2026-07-05T07:31:28Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2305.10782/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) do not differentially represent numbers, which are pervasive in text. In contrast, neuroscience research has identified distinct neural representations for numbers and words. In this work, we investigate how well popular LLMs capture the magnitudes of numbers (e.g., that $4 < 5$) from a behavioral lens. Prior research on the representational capabilities of LLMs evaluates whether they show human-level performance, for instance, high overall accuracy on standard benchmarks. Here, we ask a different question, one inspired by cognitive science: How closely do the numb","authors_text":"Khushi Bhardwaj, Raj Sanjay Shah, Reba Koenen, Sashank Varma, Vijay Marupudi","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2023-05-18T07:50:44Z","title":"Human Behavioral Benchmarking: Numeric Magnitude Comparison Effects in Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.10782","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:3be79b268901dc41ba2afb74308491355272c3519e12cf674daa9175b40059ed","target":"record","created_at":"2026-07-05T07:31:28Z","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":"8ec809073586cdb0aab2e3d5d9ac9dfce4883850b688016a82793ea2226a3a01","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2023-05-18T07:50:44Z","title_canon_sha256":"54595f2b8be874df4dfc966c15a4a3810e0c9378c533499c54436055534cba82"},"schema_version":"1.0","source":{"id":"2305.10782","kind":"arxiv","version":3}},"canonical_sha256":"e0be31c8137841c1e6ec7ebe50e87c01fb17417074a5de3c593a56b3a1e51df8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e0be31c8137841c1e6ec7ebe50e87c01fb17417074a5de3c593a56b3a1e51df8","first_computed_at":"2026-07-05T07:31:28.330630Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:31:28.330630Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"f4vbahA1SIiiSUXN449k08CkQ3z/Q3awLMhhn9VpemdwyKmS3P+8qKFCneArg+hZc+wT6AdkUb3kv7ZzLldmBg==","signature_status":"signed_v1","signed_at":"2026-07-05T07:31:28.331180Z","signed_message":"canonical_sha256_bytes"},"source_id":"2305.10782","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3be79b268901dc41ba2afb74308491355272c3519e12cf674daa9175b40059ed","sha256:0405ea7a27e688fd1a3781aca5a493f67d4d70965e47bab10b4d14b1b4bfadf1"],"state_sha256":"72b56765599cc56f7250815ccefcffda2f396ce8de21743ef9aad74010e01a2a"}