{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BB7N3S543YTJMPNDCDFCZMRIS7","short_pith_number":"pith:BB7N3S54","schema_version":"1.0","canonical_sha256":"087eddcbbcde26963da310ca2cb22897d87c84aba981a8c28fc6e26ac4655e5a","source":{"kind":"arxiv","id":"2511.02130","version":2},"attestation_state":"computed","paper":{"title":"Re-FORC: Adaptive Reward Prediction for Efficient Chain-of-Thought Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Aditya Golatkar, Alessandro Achille, Michael Kleinman, Renos Zabounidis, Stefano Soatto, Wei Xia","submitted_at":"2025-11-03T23:47:49Z","abstract_excerpt":"We propose Re-FORC, an adaptive reward prediction method that, given a query, enables prediction of the expected future rewards as a function of the number of future thinking tokens. Re-FORC trains a lightweight adapter on reasoning models, demonstrating improved prediction with longer reasoning and larger models. Re-FORC enables: 1) early stopping of unpromising reasoning chains, reducing compute by up to 26\\% compared to fixed-budget cutoffs, while maintaining accuracy, 2) optimized model and thinking length selection that outperforms the largest model alone--- reaching 1.7 percentage points"},"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":"2511.02130","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-11-03T23:47:49Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"9a6c9ac0328c5bb2307007e24f9e16f03ed8e7d8520d9e703b41b6f036f45090","abstract_canon_sha256":"029665988dac1b4d036cc7f093945236411db24ada2af51d5a2f516568cd1f0c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-27T00:19:44.997529Z","signature_b64":"sTz5e6nwl8Z4OIkNJ8VVNblu5MD/8BOoDtKHAaQPejSiserQpqFHkkBH20pQPRpxQRFtnMqAz0VI7VANkvQfCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"087eddcbbcde26963da310ca2cb22897d87c84aba981a8c28fc6e26ac4655e5a","last_reissued_at":"2026-07-27T00:19:44.996606Z","signature_status":"signed_v1","first_computed_at":"2026-07-27T00:19:44.996606Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Re-FORC: Adaptive Reward Prediction for Efficient Chain-of-Thought Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Aditya Golatkar, Alessandro Achille, Michael Kleinman, Renos Zabounidis, Stefano Soatto, Wei Xia","submitted_at":"2025-11-03T23:47:49Z","abstract_excerpt":"We propose Re-FORC, an adaptive reward prediction method that, given a query, enables prediction of the expected future rewards as a function of the number of future thinking tokens. Re-FORC trains a lightweight adapter on reasoning models, demonstrating improved prediction with longer reasoning and larger models. Re-FORC enables: 1) early stopping of unpromising reasoning chains, reducing compute by up to 26\\% compared to fixed-budget cutoffs, while maintaining accuracy, 2) optimized model and thinking length selection that outperforms the largest model alone--- reaching 1.7 percentage points"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2511.02130","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2511.02130/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":"2511.02130","created_at":"2026-07-27T00:19:44.997033+00:00"},{"alias_kind":"arxiv_version","alias_value":"2511.02130v2","created_at":"2026-07-27T00:19:44.997033+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.02130","created_at":"2026-07-27T00:19:44.997033+00:00"},{"alias_kind":"pith_short_12","alias_value":"BB7N3S543YTJ","created_at":"2026-07-27T00:19:44.997033+00:00"},{"alias_kind":"pith_short_16","alias_value":"BB7N3S543YTJMPND","created_at":"2026-07-27T00:19:44.997033+00:00"},{"alias_kind":"pith_short_8","alias_value":"BB7N3S54","created_at":"2026-07-27T00:19:44.997033+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.09741","citing_title":"ExecTune: Effective Steering of Black-Box LLMs with Guide Models","ref_index":34,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7","json":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7.json","graph_json":"https://pith.science/api/pith-number/BB7N3S543YTJMPNDCDFCZMRIS7/graph.json","events_json":"https://pith.science/api/pith-number/BB7N3S543YTJMPNDCDFCZMRIS7/events.json","paper":"https://pith.science/paper/BB7N3S54"},"agent_actions":{"view_html":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7","download_json":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7.json","view_paper":"https://pith.science/paper/BB7N3S54","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2511.02130&json=true","fetch_graph":"https://pith.science/api/pith-number/BB7N3S543YTJMPNDCDFCZMRIS7/graph.json","fetch_events":"https://pith.science/api/pith-number/BB7N3S543YTJMPNDCDFCZMRIS7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7/action/storage_attestation","attest_author":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7/action/author_attestation","sign_citation":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7/action/citation_signature","submit_replication":"https://pith.science/pith/BB7N3S543YTJMPNDCDFCZMRIS7/action/replication_record"}},"created_at":"2026-07-27T00:19:44.997033+00:00","updated_at":"2026-07-27T00:19:44.997033+00:00"}