{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:ZAS3BS3GDTMNZ5ZXS6R5MEBPDL","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":"76d46be4c8afba9285feb51ab901a14aadad1d7005282057a35cd832bd5f7fe1","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-12T17:51:10Z","title_canon_sha256":"385c9513dad506323bf4f2f02194124608fe722dfcce72a08bda3cd89964adf4"},"schema_version":"1.0","source":{"id":"2310.08559","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.08559","created_at":"2026-07-05T08:21:41Z"},{"alias_kind":"arxiv_version","alias_value":"2310.08559v4","created_at":"2026-07-05T08:21:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.08559","created_at":"2026-07-05T08:21:41Z"},{"alias_kind":"pith_short_12","alias_value":"ZAS3BS3GDTMN","created_at":"2026-07-05T08:21:41Z"},{"alias_kind":"pith_short_16","alias_value":"ZAS3BS3GDTMNZ5ZX","created_at":"2026-07-05T08:21:41Z"},{"alias_kind":"pith_short_8","alias_value":"ZAS3BS3G","created_at":"2026-07-05T08:21:41Z"}],"graph_snapshots":[{"event_id":"sha256:f2119617a8d96ec076325f271b3988f3805b8d3b303a12eed7cbd8c79e869ab4","target":"graph","created_at":"2026-07-05T08:21:41Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2310.08559/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The ability to derive underlying principles from a handful of observations and then generalize to novel situations -- known as inductive reasoning -- is central to human intelligence. Prior work suggests that language models (LMs) often fall short on inductive reasoning, despite achieving impressive success on research benchmarks. In this work, we conduct a systematic study of the inductive reasoning capabilities of LMs through iterative hypothesis refinement, a technique that more closely mirrors the human inductive process than standard input-output prompting. Iterative hypothesis refinement","authors_text":"Bailin Wang, Chandra Bhagavatula, Linlu Qiu, Liwei Jiang, Melanie Sclar, Nouha Dziri, Valentina Pyatkin, Xiang Ren, Ximing Lu, Yejin Choi, Yoon Kim","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-12T17:51:10Z","title":"Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.08559","kind":"arxiv","version":4},"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:e408dbfb0d4a87a90274428f6a4442c23a6647d11f70afc3619ae6818b099e45","target":"record","created_at":"2026-07-05T08:21:41Z","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":"76d46be4c8afba9285feb51ab901a14aadad1d7005282057a35cd832bd5f7fe1","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-12T17:51:10Z","title_canon_sha256":"385c9513dad506323bf4f2f02194124608fe722dfcce72a08bda3cd89964adf4"},"schema_version":"1.0","source":{"id":"2310.08559","kind":"arxiv","version":4}},"canonical_sha256":"c825b0cb661cd8dcf73797a3d6102f1ac6899c73f3fdfa805efb039924e42294","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c825b0cb661cd8dcf73797a3d6102f1ac6899c73f3fdfa805efb039924e42294","first_computed_at":"2026-07-05T08:21:41.940917Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:21:41.940917Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"683ydPGsv6zcs2nWBLX9cWTT1cWDL8Wvt6ucABh+uwuQ4NEX1+885X0gEh7MWvo9n6in7zY8bEVCEo+7OOOsDg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:21:41.941372Z","signed_message":"canonical_sha256_bytes"},"source_id":"2310.08559","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e408dbfb0d4a87a90274428f6a4442c23a6647d11f70afc3619ae6818b099e45","sha256:f2119617a8d96ec076325f271b3988f3805b8d3b303a12eed7cbd8c79e869ab4"],"state_sha256":"6ab689d55736c1c24ed0f6ac738bc52a0072b9e7ffa962db2bef350a34cd319b"}