{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:CNEKXVJ67O3P7T73CHXRH2RCSN","short_pith_number":"pith:CNEKXVJ6","schema_version":"1.0","canonical_sha256":"1348abd53efbb6ffcffb11ef13ea22934fc347e92d7517250060e8cfe59d86ce","source":{"kind":"arxiv","id":"2606.02248","version":1},"attestation_state":"computed","paper":{"title":"Geometric Latent Reasoning Induces Shorter Generations in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andrea Cavallaro, Ina Kodrasi, Petr Motlicek, Shashi Kumar, Yacouba Kaloga","submitted_at":"2026-06-01T13:40:55Z","abstract_excerpt":"Large language models solve complex problems by generating lengthy chains of explicit reasoning tokens. While effective, this makes reasoning expensive, length-sensitive, and constrained to (discrete) natural language. While latent reasoning offers a continuous alternative, determining useful structures for intermediate latent states is an open challenge. In this paper, we formulate latent reasoning as a geometric path-approximation problem within the model's pretrained token-embedding space. We introduce Geometric Latent Reasoning (GLR), which uses a lightweight transition head to predict ite"},"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":"2606.02248","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-01T13:40:55Z","cross_cats_sorted":[],"title_canon_sha256":"9d563747db845787ba7ab07fa1c6114a422dd0ef6eeac6e916e849070f7b4529","abstract_canon_sha256":"a68dfba4c1c01b5b09da779bd7a24a063796b00d8ec7fd4d218694757a85d3a6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T03:04:54.101160Z","signature_b64":"hzOra4ptHhElopkvrdMszpiw+EZr/QT/QAyg7wBZnZLTUBnkkVadhptV0eR8WVw8PcxrEMU0/tBZ9UwDfAFIDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1348abd53efbb6ffcffb11ef13ea22934fc347e92d7517250060e8cfe59d86ce","last_reissued_at":"2026-06-02T03:04:54.100748Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T03:04:54.100748Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Geometric Latent Reasoning Induces Shorter Generations in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andrea Cavallaro, Ina Kodrasi, Petr Motlicek, Shashi Kumar, Yacouba Kaloga","submitted_at":"2026-06-01T13:40:55Z","abstract_excerpt":"Large language models solve complex problems by generating lengthy chains of explicit reasoning tokens. While effective, this makes reasoning expensive, length-sensitive, and constrained to (discrete) natural language. While latent reasoning offers a continuous alternative, determining useful structures for intermediate latent states is an open challenge. In this paper, we formulate latent reasoning as a geometric path-approximation problem within the model's pretrained token-embedding space. We introduce Geometric Latent Reasoning (GLR), which uses a lightweight transition head to predict ite"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02248","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/2606.02248/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":"2606.02248","created_at":"2026-06-02T03:04:54.100809+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02248v1","created_at":"2026-06-02T03:04:54.100809+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02248","created_at":"2026-06-02T03:04:54.100809+00:00"},{"alias_kind":"pith_short_12","alias_value":"CNEKXVJ67O3P","created_at":"2026-06-02T03:04:54.100809+00:00"},{"alias_kind":"pith_short_16","alias_value":"CNEKXVJ67O3P7T73","created_at":"2026-06-02T03:04:54.100809+00:00"},{"alias_kind":"pith_short_8","alias_value":"CNEKXVJ6","created_at":"2026-06-02T03:04:54.100809+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/CNEKXVJ67O3P7T73CHXRH2RCSN","json":"https://pith.science/pith/CNEKXVJ67O3P7T73CHXRH2RCSN.json","graph_json":"https://pith.science/api/pith-number/CNEKXVJ67O3P7T73CHXRH2RCSN/graph.json","events_json":"https://pith.science/api/pith-number/CNEKXVJ67O3P7T73CHXRH2RCSN/events.json","paper":"https://pith.science/paper/CNEKXVJ6"},"agent_actions":{"view_html":"https://pith.science/pith/CNEKXVJ67O3P7T73CHXRH2RCSN","download_json":"https://pith.science/pith/CNEKXVJ67O3P7T73CHXRH2RCSN.json","view_paper":"https://pith.science/paper/CNEKXVJ6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02248&json=true","fetch_graph":"https://pith.science/api/pith-number/CNEKXVJ67O3P7T73CHXRH2RCSN/graph.json","fetch_events":"https://pith.science/api/pith-number/CNEKXVJ67O3P7T73CHXRH2RCSN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CNEKXVJ67O3P7T73CHXRH2RCSN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CNEKXVJ67O3P7T73CHXRH2RCSN/action/storage_attestation","attest_author":"https://pith.science/pith/CNEKXVJ67O3P7T73CHXRH2RCSN/action/author_attestation","sign_citation":"https://pith.science/pith/CNEKXVJ67O3P7T73CHXRH2RCSN/action/citation_signature","submit_replication":"https://pith.science/pith/CNEKXVJ67O3P7T73CHXRH2RCSN/action/replication_record"}},"created_at":"2026-06-02T03:04:54.100809+00:00","updated_at":"2026-06-02T03:04:54.100809+00:00"}