{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:KFNEXE427NDOHBEOVUOAZ3CDD4","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"ff6f48d22cf417a155f6f9c74452f2ba17ea6a4cedea10ed547be974b9db1781","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-12-08T11:39:43Z","title_canon_sha256":"65f95b024d0449dc5a4d9152aa44ed7c9baf0699060d8b575a2e2585e4b444f2"},"schema_version":"1.0","source":{"id":"2512.07461","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2512.07461","created_at":"2026-05-17T23:39:00Z"},{"alias_kind":"arxiv_version","alias_value":"2512.07461v3","created_at":"2026-05-17T23:39:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.07461","created_at":"2026-05-17T23:39:00Z"},{"alias_kind":"pith_short_12","alias_value":"KFNEXE427NDO","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"KFNEXE427NDOHBEO","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"KFNEXE42","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:d70216900d6a2f2f8ba39c8c82e0b82e7a055849ebcbc3d1f0415be439ec9c7f","target":"graph","created_at":"2026-05-17T23:39:00Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"NPR trained on Qwen3-4B achieves performance gains of up to 24.5% and inference speedups up to 4.6x. Unlike prior baselines that often fall back to autoregressive decoding, NPR demonstrates 100% genuine parallel execution."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The self-distilled progressive training paradigm successfully transitions the model to native parallel cognition with strict topological constraints without external supervision or falling back to sequential behavior."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"NPR trains LLMs to reason in parallel via self-distilled RL, delivering up to 24.5% performance gains and 4.6x speedups with 100% genuine parallel execution on reasoning benchmarks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Large language models can learn genuine parallel reasoning on their own through self-distilled reinforcement learning."}],"snapshot_sha256":"214879fc52049511adfe68c891e7a14efd86ff8859357ae9ba3aafb7195fa9a3"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"We introduce Native Parallel Reasoner (NPR), a teacher-free framework that enables Large Language Models (LLMs) to self-evolve genuine parallel reasoning capabilities. NPR transforms the model from sequential emulation to native parallel cognition through three key innovations: 1) a self-distilled progressive training paradigm that transitions from ``cold-start'' format discovery to strict topological constraints without external supervision; 2) a novel Parallel-Aware Policy Optimization (PAPO) algorithm that optimizes branching policies directly within the execution graph, allowing the model ","authors_text":"Jun Bai, Shuyi Zhang, Song-Chun Zhu, Tong Wu, Yang Liu, Yanting Wang, Zilong Zheng, Zixia Jia, Ziyong Lin","cross_cats":[],"headline":"Large language models can learn genuine parallel reasoning on their own through self-distilled reinforcement learning.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-12-08T11:39:43Z","title":"Native Parallel Reasoner: Reasoning in Parallelism via Self-Distilled Reinforcement Learning"},"references":{"count":25,"internal_anchors":7,"resolved_work":25,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Doing: Agents that Reason by Scaling Test-Time Interaction , author=","work_id":"1619f6ef-5a19-4b4e-9731-fef04ca5be72","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Multiverse: Your language models secretly decide how to parallelize and merge generation","work_id":"292f97a6-b3ce-43b7-9296-047c8f9fecc0","year":null},{"cited_arxiv_id":"2402.03300","doi":"","is_internal_anchor":true,"ref_index":3,"title":"DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models","work_id":"c5006563-f3ec-438a-9e35-b7b484f34828","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Parallel-r1: Towards parallel thinking via reinforcement learning","work_id":"75a4a861-4da2-4147-926b-d361952ab5e5","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Parallelsearch: Train your llms to decompose query and search sub-queries in parallel with reinforcement learning","work_id":"1a767f87-f215-4e70-bdcd-a05327fe1314","year":null}],"snapshot_sha256":"627a65004213ae8cf3c79163fffb511675df75247d3870b24adefa23a2e6ff79"},"source":{"id":"2512.07461","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-17T01:05:32.222004Z","id":"3b615a40-54df-42d0-9d2d-50cce85421b8","model_set":{"reader":"grok-4.3"},"one_line_summary":"NPR trains LLMs to reason in parallel via self-distilled RL, delivering up to 24.5% performance gains and 4.6x speedups with 100% genuine parallel execution on reasoning benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Large language models can learn genuine parallel reasoning on their own through self-distilled reinforcement learning.","strongest_claim":"NPR trained on Qwen3-4B achieves performance gains of up to 24.5% and inference speedups up to 4.6x. Unlike prior baselines that often fall back to autoregressive decoding, NPR demonstrates 100% genuine parallel execution.","weakest_assumption":"The self-distilled progressive training paradigm successfully transitions the model to native parallel cognition with strict topological constraints without external supervision or falling back to sequential behavior."}},"verdict_id":"3b615a40-54df-42d0-9d2d-50cce85421b8"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6883414069ff8de91475162249ebb215279d33c83cd401abbdbcf5e16a2b4bf1","target":"record","created_at":"2026-05-17T23:39:00Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"ff6f48d22cf417a155f6f9c74452f2ba17ea6a4cedea10ed547be974b9db1781","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-12-08T11:39:43Z","title_canon_sha256":"65f95b024d0449dc5a4d9152aa44ed7c9baf0699060d8b575a2e2585e4b444f2"},"schema_version":"1.0","source":{"id":"2512.07461","kind":"arxiv","version":3}},"canonical_sha256":"515a4b939afb46e3848ead1c0cec431f171c8961ac396fc7888cc25693eca8b2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"515a4b939afb46e3848ead1c0cec431f171c8961ac396fc7888cc25693eca8b2","first_computed_at":"2026-05-17T23:39:00.571222Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:00.571222Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Qco0agCh0xUhgDhwH/cqvuRKrCuIsiFqZCRRgoP0n5KZUhuUcNbqbfCm1YbQHVRMl3hu+5IVLV4TZHVNE5fyBg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:00.572054Z","signed_message":"canonical_sha256_bytes"},"source_id":"2512.07461","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6883414069ff8de91475162249ebb215279d33c83cd401abbdbcf5e16a2b4bf1","sha256:d70216900d6a2f2f8ba39c8c82e0b82e7a055849ebcbc3d1f0415be439ec9c7f"],"state_sha256":"42af2b8712ddfb85ab778401f8969deaec213930b7f6768b8bcce37c68770176"}