{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:6SOSWDQSUUWAQVPMRRY4PFIPNP","short_pith_number":"pith:6SOSWDQS","schema_version":"1.0","canonical_sha256":"f49d2b0e12a52c0855ec8c71c7950f6bdb88671339d75d4a30c5267d29a0786d","source":{"kind":"arxiv","id":"2506.09890","version":1},"attestation_state":"computed","paper":{"title":"The Emergence of Abstract Thought in Large Language Models Beyond Any Language","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"An Zhang, Junnan Li, Kenji Kawaguchi, Michael Qizhe Shieh, Shafiq Joty, Tat-Seng Chua, Wenxuan Zhang, Yang Zhang, Yiran Zhao, Yuxin Chen","submitted_at":"2025-06-11T16:00:54Z","abstract_excerpt":"As large language models (LLMs) continue to advance, their capacity to function effectively across a diverse range of languages has shown marked improvement. Preliminary studies observe that the hidden activations of LLMs often resemble English, even when responding to non-English prompts. This has led to the widespread assumption that LLMs may \"think\" in English. However, more recent results showing strong multilingual performance, even surpassing English performance on specific tasks in other languages, challenge this view. In this work, we find that LLMs progressively develop a core languag"},"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":"2506.09890","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-11T16:00:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c764e1cf2334cb9face8cd7cb91ffd24c3df50b2dc7db8a6d6875df18399f1cc","abstract_canon_sha256":"975c075ba5a275f0a73d6168f50069efb0b5065407dbd1d0777f6b43d7a7c284"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:19:58.368850Z","signature_b64":"sZr6an4l+AedDN8kEvZpSMb+NTcEHm33/1dUFsuO3R5a1r84Y26Sq9WFGqhwfFaHTcIhL6/q620sbszYCcuoBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f49d2b0e12a52c0855ec8c71c7950f6bdb88671339d75d4a30c5267d29a0786d","last_reissued_at":"2026-07-05T11:19:58.368350Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:19:58.368350Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Emergence of Abstract Thought in Large Language Models Beyond Any Language","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"An Zhang, Junnan Li, Kenji Kawaguchi, Michael Qizhe Shieh, Shafiq Joty, Tat-Seng Chua, Wenxuan Zhang, Yang Zhang, Yiran Zhao, Yuxin Chen","submitted_at":"2025-06-11T16:00:54Z","abstract_excerpt":"As large language models (LLMs) continue to advance, their capacity to function effectively across a diverse range of languages has shown marked improvement. Preliminary studies observe that the hidden activations of LLMs often resemble English, even when responding to non-English prompts. This has led to the widespread assumption that LLMs may \"think\" in English. However, more recent results showing strong multilingual performance, even surpassing English performance on specific tasks in other languages, challenge this view. In this work, we find that LLMs progressively develop a core languag"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.09890","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/2506.09890/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":"2506.09890","created_at":"2026-07-05T11:19:58.368413+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.09890v1","created_at":"2026-07-05T11:19:58.368413+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.09890","created_at":"2026-07-05T11:19:58.368413+00:00"},{"alias_kind":"pith_short_12","alias_value":"6SOSWDQSUUWA","created_at":"2026-07-05T11:19:58.368413+00:00"},{"alias_kind":"pith_short_16","alias_value":"6SOSWDQSUUWAQVPM","created_at":"2026-07-05T11:19:58.368413+00:00"},{"alias_kind":"pith_short_8","alias_value":"6SOSWDQS","created_at":"2026-07-05T11:19:58.368413+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.21303","citing_title":"From Circuit Evidence to Mechanistic Theory: An Inductive Logic Approach","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08881","citing_title":"Targeted Interpretable Safety Neuron Enhancement for Multilingual Vision-Language Large Models","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP","json":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP.json","graph_json":"https://pith.science/api/pith-number/6SOSWDQSUUWAQVPMRRY4PFIPNP/graph.json","events_json":"https://pith.science/api/pith-number/6SOSWDQSUUWAQVPMRRY4PFIPNP/events.json","paper":"https://pith.science/paper/6SOSWDQS"},"agent_actions":{"view_html":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP","download_json":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP.json","view_paper":"https://pith.science/paper/6SOSWDQS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.09890&json=true","fetch_graph":"https://pith.science/api/pith-number/6SOSWDQSUUWAQVPMRRY4PFIPNP/graph.json","fetch_events":"https://pith.science/api/pith-number/6SOSWDQSUUWAQVPMRRY4PFIPNP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP/action/storage_attestation","attest_author":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP/action/author_attestation","sign_citation":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP/action/citation_signature","submit_replication":"https://pith.science/pith/6SOSWDQSUUWAQVPMRRY4PFIPNP/action/replication_record"}},"created_at":"2026-07-05T11:19:58.368413+00:00","updated_at":"2026-07-05T11:19:58.368413+00:00"}