{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CVFYCLPVTCHS4NW3DGY2XA6UJW","short_pith_number":"pith:CVFYCLPV","schema_version":"1.0","canonical_sha256":"154b812df5988f2e36db19b1ab83d44daf7a7fd842d73b2a67caa94b9ece27e0","source":{"kind":"arxiv","id":"2506.23274","version":4},"attestation_state":"computed","paper":{"title":"Real-Time Progress Prediction in Reasoning Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Anders S{\\o}gaard, Constanza Fierro, Hans Peter Lyngs{\\o}e Raaschou-Jensen","submitted_at":"2025-06-29T15:01:01Z","abstract_excerpt":"Recent reasoning language models, particularly those that employ long latent chains of thought, achieve strong performance on complex agentic tasks. However, as these models operate over increasingly long time horizons, their internal progress becomes opaque to users, making expectation management and real-time oversight difficult. In this work, we investigate whether real-time progress prediction is feasible for such models. We first test whether hidden states encode progress information by discretizing reasoning trajectories and training a linear probe to classify reasoning states. We then f"},"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":"2506.23274","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-29T15:01:01Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"63be4b601aa260598b10ddaa6d865fa0c2f96e1ae19aa2b8f09418e5ac0601d9","abstract_canon_sha256":"a46128a41e87dcafa1af28e986a47aabfef6b354f86d92421cd777a76db8631f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:05:35.425655Z","signature_b64":"6HGwW0qN5t+OyoyAyOTb0fuXCfh73Yeqx22AVmXr2w3nEf1ltiKaRqBEPJQNjjH2k3B20VPHrugEi76MZxL/Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"154b812df5988f2e36db19b1ab83d44daf7a7fd842d73b2a67caa94b9ece27e0","last_reissued_at":"2026-05-27T01:05:35.425089Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:05:35.425089Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Real-Time Progress Prediction in Reasoning Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Anders S{\\o}gaard, Constanza Fierro, Hans Peter Lyngs{\\o}e Raaschou-Jensen","submitted_at":"2025-06-29T15:01:01Z","abstract_excerpt":"Recent reasoning language models, particularly those that employ long latent chains of thought, achieve strong performance on complex agentic tasks. However, as these models operate over increasingly long time horizons, their internal progress becomes opaque to users, making expectation management and real-time oversight difficult. In this work, we investigate whether real-time progress prediction is feasible for such models. We first test whether hidden states encode progress information by discretizing reasoning trajectories and training a linear probe to classify reasoning states. We then f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.23274","kind":"arxiv","version":4},"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/2506.23274/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":"2506.23274","created_at":"2026-05-27T01:05:35.425150+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.23274v4","created_at":"2026-05-27T01:05:35.425150+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.23274","created_at":"2026-05-27T01:05:35.425150+00:00"},{"alias_kind":"pith_short_12","alias_value":"CVFYCLPVTCHS","created_at":"2026-05-27T01:05:35.425150+00:00"},{"alias_kind":"pith_short_16","alias_value":"CVFYCLPVTCHS4NW3","created_at":"2026-05-27T01:05:35.425150+00:00"},{"alias_kind":"pith_short_8","alias_value":"CVFYCLPV","created_at":"2026-05-27T01:05:35.425150+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.05851","citing_title":"Hypothesis generation and updating in large language models","ref_index":15,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW","json":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW.json","graph_json":"https://pith.science/api/pith-number/CVFYCLPVTCHS4NW3DGY2XA6UJW/graph.json","events_json":"https://pith.science/api/pith-number/CVFYCLPVTCHS4NW3DGY2XA6UJW/events.json","paper":"https://pith.science/paper/CVFYCLPV"},"agent_actions":{"view_html":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW","download_json":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW.json","view_paper":"https://pith.science/paper/CVFYCLPV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.23274&json=true","fetch_graph":"https://pith.science/api/pith-number/CVFYCLPVTCHS4NW3DGY2XA6UJW/graph.json","fetch_events":"https://pith.science/api/pith-number/CVFYCLPVTCHS4NW3DGY2XA6UJW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW/action/storage_attestation","attest_author":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW/action/author_attestation","sign_citation":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW/action/citation_signature","submit_replication":"https://pith.science/pith/CVFYCLPVTCHS4NW3DGY2XA6UJW/action/replication_record"}},"created_at":"2026-05-27T01:05:35.425150+00:00","updated_at":"2026-05-27T01:05:35.425150+00:00"}