{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QMZQ7JO65NSKD43PWHV2SATIJJ","short_pith_number":"pith:QMZQ7JO6","schema_version":"1.0","canonical_sha256":"83330fa5deeb64a1f36fb1eba902684a48bb2f850d3cbda4686dc1c17f5d5e9c","source":{"kind":"arxiv","id":"2506.11274","version":2},"attestation_state":"computed","paper":{"title":"Learning a Continue-Thinking Token for Enhanced Test-Time Scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Elad Tolochinsky, Liran Ringel, Yaniv Romano","submitted_at":"2025-06-12T20:28:54Z","abstract_excerpt":"Test-time scaling has emerged as an effective approach for improving language model performance by utilizing additional compute at inference time. Recent studies have shown that overriding end-of-thinking tokens (e.g., replacing \"</think>\" with \"Wait\") can extend reasoning steps and improve accuracy. In this work, we explore whether a dedicated continue-thinking token can be learned to trigger extended reasoning. We augment a distilled version of DeepSeek-R1 with a single learned \"<|continue-thinking|>\" token, training only its embedding via reinforcement learning while keeping the model weigh"},"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":"2506.11274","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-06-12T20:28:54Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"dbc1164693ad61a4e369d94b91902fa86475815f06cd4e5d864550001845016d","abstract_canon_sha256":"8e1f3235ff71e4f21336e7c775834a0d6963edfb0f8f226886ab05de284a0070"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:10:12.579242Z","signature_b64":"D8gPjtnZSM3cvgcsJnhS7TrIOqUSWJGGkXqijC4w3L+fO2MB1h2phTkISRudOwviXQk85gQE0XYXk8STmcn/Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"83330fa5deeb64a1f36fb1eba902684a48bb2f850d3cbda4686dc1c17f5d5e9c","last_reissued_at":"2026-05-18T03:10:12.578778Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:10:12.578778Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning a Continue-Thinking Token for Enhanced Test-Time Scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Elad Tolochinsky, Liran Ringel, Yaniv Romano","submitted_at":"2025-06-12T20:28:54Z","abstract_excerpt":"Test-time scaling has emerged as an effective approach for improving language model performance by utilizing additional compute at inference time. Recent studies have shown that overriding end-of-thinking tokens (e.g., replacing \"</think>\" with \"Wait\") can extend reasoning steps and improve accuracy. In this work, we explore whether a dedicated continue-thinking token can be learned to trigger extended reasoning. We augment a distilled version of DeepSeek-R1 with a single learned \"<|continue-thinking|>\" token, training only its embedding via reinforcement learning while keeping the model weigh"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.11274","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"2506.11274","created_at":"2026-05-18T03:10:12.578849+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.11274v2","created_at":"2026-05-18T03:10:12.578849+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.11274","created_at":"2026-05-18T03:10:12.578849+00:00"},{"alias_kind":"pith_short_12","alias_value":"QMZQ7JO65NSK","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"QMZQ7JO65NSKD43P","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"QMZQ7JO6","created_at":"2026-05-18T12:33:37.589309+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/QMZQ7JO65NSKD43PWHV2SATIJJ","json":"https://pith.science/pith/QMZQ7JO65NSKD43PWHV2SATIJJ.json","graph_json":"https://pith.science/api/pith-number/QMZQ7JO65NSKD43PWHV2SATIJJ/graph.json","events_json":"https://pith.science/api/pith-number/QMZQ7JO65NSKD43PWHV2SATIJJ/events.json","paper":"https://pith.science/paper/QMZQ7JO6"},"agent_actions":{"view_html":"https://pith.science/pith/QMZQ7JO65NSKD43PWHV2SATIJJ","download_json":"https://pith.science/pith/QMZQ7JO65NSKD43PWHV2SATIJJ.json","view_paper":"https://pith.science/paper/QMZQ7JO6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.11274&json=true","fetch_graph":"https://pith.science/api/pith-number/QMZQ7JO65NSKD43PWHV2SATIJJ/graph.json","fetch_events":"https://pith.science/api/pith-number/QMZQ7JO65NSKD43PWHV2SATIJJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QMZQ7JO65NSKD43PWHV2SATIJJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QMZQ7JO65NSKD43PWHV2SATIJJ/action/storage_attestation","attest_author":"https://pith.science/pith/QMZQ7JO65NSKD43PWHV2SATIJJ/action/author_attestation","sign_citation":"https://pith.science/pith/QMZQ7JO65NSKD43PWHV2SATIJJ/action/citation_signature","submit_replication":"https://pith.science/pith/QMZQ7JO65NSKD43PWHV2SATIJJ/action/replication_record"}},"created_at":"2026-05-18T03:10:12.578849+00:00","updated_at":"2026-05-18T03:10:12.578849+00:00"}