{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:LZYS5Y2NGOD76NSOMOLEKW2HQ4","short_pith_number":"pith:LZYS5Y2N","canonical_record":{"source":{"id":"2604.16565","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-17T10:17:16Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b3efce7e3030e59293df698065a15bb60510f6d3c385101269500421119326cf","abstract_canon_sha256":"ca0d68b4cc54b8bef24b861f412c993ea74f64ec3962751313e1980d92074ca4"},"schema_version":"1.0"},"canonical_sha256":"5e712ee34d3387ff364e6396455b47873837b572eb67bb677b205690ee447ad1","source":{"kind":"arxiv","id":"2604.16565","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.16565","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"arxiv_version","alias_value":"2604.16565v3","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.16565","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"pith_short_12","alias_value":"LZYS5Y2NGOD7","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"pith_short_16","alias_value":"LZYS5Y2NGOD76NSO","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"pith_short_8","alias_value":"LZYS5Y2N","created_at":"2026-05-28T02:04:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:LZYS5Y2NGOD76NSOMOLEKW2HQ4","target":"record","payload":{"canonical_record":{"source":{"id":"2604.16565","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-17T10:17:16Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b3efce7e3030e59293df698065a15bb60510f6d3c385101269500421119326cf","abstract_canon_sha256":"ca0d68b4cc54b8bef24b861f412c993ea74f64ec3962751313e1980d92074ca4"},"schema_version":"1.0"},"canonical_sha256":"5e712ee34d3387ff364e6396455b47873837b572eb67bb677b205690ee447ad1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T02:04:47.981257Z","signature_b64":"m8184IUm+ylxD+xJ0iMdj97Q5zrBITRmmf3HtmOoLCF+GihCSn7NB05lpoP+YpwzJOK0UjM/SQHUQaf/UHqGCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5e712ee34d3387ff364e6396455b47873837b572eb67bb677b205690ee447ad1","last_reissued_at":"2026-05-28T02:04:47.980747Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T02:04:47.980747Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.16565","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-28T02:04:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cIEtMvlUZS2HQnrePP5x2Ddo8+th9PB2zipYnO14VV9E1JDnVXHOwFQKVUuTQ0njJoXbzLDnQe67kOe7X0f0Bw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-09T12:34:54.729673Z"},"content_sha256":"43c188e5ef7d71766dd1b673f8b617379931b408691066a6e1c593211d232dce","schema_version":"1.0","event_id":"sha256:43c188e5ef7d71766dd1b673f8b617379931b408691066a6e1c593211d232dce"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:LZYS5Y2NGOD76NSOMOLEKW2HQ4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Reasoning on the Manifold: Bidirectional Consistency for Self-Verification in Diffusion Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Diffusion language models can verify their own reasoning by measuring how stable generated sequences remain under a forward-masking and backward-reconstruction cycle on the learned manifold.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Guanghao Li, Hengyu Zeng, Jian Pu, Jiaoyang Ruan, Jie Fu, Liang Du, Xin Gao, Yinda Chen","submitted_at":"2026-04-17T10:17:16Z","abstract_excerpt":"While Diffusion Large Language Models (dLLMs) offer structural advantages for global planning, efficiently verifying that they arrive at correct answers via valid reasoning traces remains a critical challenge. In this work, we propose a geometric perspective: Reasoning on the Manifold. We hypothesize that valid generation trajectories reside as stable attractors on the high-density manifold of the learned distribution, whereas invalid paths exhibit off-manifold drift. To operationalize this, we introduce Bidirectional Manifold Consistency (BMC), a training-free, unsupervised metric that quanti"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our results establish intrinsic geometric stability as a robust indicator of correctness for dLLMs.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"Valid generation trajectories reside as stable attractors on the high-density manifold of the learned distribution, and the forward-masking backward-reconstruction cycle accurately quantifies this stability as a proxy for correctness.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Bidirectional Manifold Consistency (BMC) is a geometric, training-free metric that quantifies stability of reasoning trajectories in diffusion LLMs to enable self-verification, rejection sampling, and alignment without ground truth.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Diffusion language models can verify their own reasoning by measuring how stable generated sequences remain under a forward-masking and backward-reconstruction cycle on the learned manifold.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f3c344123a4bed2cb2f1935ce15b3a193caac0151ee6211307b27f51c9ef361c"},"source":{"id":"2604.16565","kind":"arxiv","version":3},"verdict":{"id":"9e1bc318-cdf3-4987-8c35-2bf311695f33","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T08:42:15.048118Z","strongest_claim":"Our results establish intrinsic geometric stability as a robust indicator of correctness for dLLMs.","one_line_summary":"Bidirectional Manifold Consistency (BMC) is a geometric, training-free metric that quantifies stability of reasoning trajectories in diffusion LLMs to enable self-verification, rejection sampling, and alignment without ground truth.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"Valid generation trajectories reside as stable attractors on the high-density manifold of the learned distribution, and the forward-masking backward-reconstruction cycle accurately quantifies this stability as a proxy for correctness.","pith_extraction_headline":"Diffusion language models can verify their own reasoning by measuring how stable generated sequences remain under a forward-masking and backward-reconstruction cycle on the learned manifold."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.16565/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"},"verdict_id":"9e1bc318-cdf3-4987-8c35-2bf311695f33"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-28T02:04:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rTgo1KaGZ4loAyMCU4l4LyNDYcUaLCBomXI7zDequpAHDX8Caj0u60Xmaaf+FuM2gRfvIpAf/mydnTnIcyQaCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-09T12:34:54.730129Z"},"content_sha256":"0d16ec5f1ef963bc03f3f1849e6445a693f0080eef89778870d5f7ea0fb2b746","schema_version":"1.0","event_id":"sha256:0d16ec5f1ef963bc03f3f1849e6445a693f0080eef89778870d5f7ea0fb2b746"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LZYS5Y2NGOD76NSOMOLEKW2HQ4/bundle.json","state_url":"https://pith.science/pith/LZYS5Y2NGOD76NSOMOLEKW2HQ4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LZYS5Y2NGOD76NSOMOLEKW2HQ4/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-09T12:34:54Z","links":{"resolver":"https://pith.science/pith/LZYS5Y2NGOD76NSOMOLEKW2HQ4","bundle":"https://pith.science/pith/LZYS5Y2NGOD76NSOMOLEKW2HQ4/bundle.json","state":"https://pith.science/pith/LZYS5Y2NGOD76NSOMOLEKW2HQ4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LZYS5Y2NGOD76NSOMOLEKW2HQ4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LZYS5Y2NGOD76NSOMOLEKW2HQ4","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":"ca0d68b4cc54b8bef24b861f412c993ea74f64ec3962751313e1980d92074ca4","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-17T10:17:16Z","title_canon_sha256":"b3efce7e3030e59293df698065a15bb60510f6d3c385101269500421119326cf"},"schema_version":"1.0","source":{"id":"2604.16565","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.16565","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"arxiv_version","alias_value":"2604.16565v3","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.16565","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"pith_short_12","alias_value":"LZYS5Y2NGOD7","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"pith_short_16","alias_value":"LZYS5Y2NGOD76NSO","created_at":"2026-05-28T02:04:47Z"},{"alias_kind":"pith_short_8","alias_value":"LZYS5Y2N","created_at":"2026-05-28T02:04:47Z"}],"graph_snapshots":[{"event_id":"sha256:0d16ec5f1ef963bc03f3f1849e6445a693f0080eef89778870d5f7ea0fb2b746","target":"graph","created_at":"2026-05-28T02:04:47Z","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":"Our results establish intrinsic geometric stability as a robust indicator of correctness for dLLMs."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"Valid generation trajectories reside as stable attractors on the high-density manifold of the learned distribution, and the forward-masking backward-reconstruction cycle accurately quantifies this stability as a proxy for correctness."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Bidirectional Manifold Consistency (BMC) is a geometric, training-free metric that quantifies stability of reasoning trajectories in diffusion LLMs to enable self-verification, rejection sampling, and alignment without ground truth."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Diffusion language models can verify their own reasoning by measuring how stable generated sequences remain under a forward-masking and backward-reconstruction cycle on the learned manifold."}],"snapshot_sha256":"f3c344123a4bed2cb2f1935ce15b3a193caac0151ee6211307b27f51c9ef361c"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.16565/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"While Diffusion Large Language Models (dLLMs) offer structural advantages for global planning, efficiently verifying that they arrive at correct answers via valid reasoning traces remains a critical challenge. In this work, we propose a geometric perspective: Reasoning on the Manifold. We hypothesize that valid generation trajectories reside as stable attractors on the high-density manifold of the learned distribution, whereas invalid paths exhibit off-manifold drift. To operationalize this, we introduce Bidirectional Manifold Consistency (BMC), a training-free, unsupervised metric that quanti","authors_text":"Guanghao Li, Hengyu Zeng, Jian Pu, Jiaoyang Ruan, Jie Fu, Liang Du, Xin Gao, Yinda Chen","cross_cats":["cs.AI"],"headline":"Diffusion language models can verify their own reasoning by measuring how stable generated sequences remain under a forward-masking and backward-reconstruction cycle on the learned manifold.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-17T10:17:16Z","title":"Reasoning on the Manifold: Bidirectional Consistency for Self-Verification in Diffusion Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.16565","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-10T08:42:15.048118Z","id":"9e1bc318-cdf3-4987-8c35-2bf311695f33","model_set":{"reader":"grok-4.3"},"one_line_summary":"Bidirectional Manifold Consistency (BMC) is a geometric, training-free metric that quantifies stability of reasoning trajectories in diffusion LLMs to enable self-verification, rejection sampling, and alignment without ground truth.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Diffusion language models can verify their own reasoning by measuring how stable generated sequences remain under a forward-masking and backward-reconstruction cycle on the learned manifold.","strongest_claim":"Our results establish intrinsic geometric stability as a robust indicator of correctness for dLLMs.","weakest_assumption":"Valid generation trajectories reside as stable attractors on the high-density manifold of the learned distribution, and the forward-masking backward-reconstruction cycle accurately quantifies this stability as a proxy for correctness."}},"verdict_id":"9e1bc318-cdf3-4987-8c35-2bf311695f33"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:43c188e5ef7d71766dd1b673f8b617379931b408691066a6e1c593211d232dce","target":"record","created_at":"2026-05-28T02:04:47Z","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":"ca0d68b4cc54b8bef24b861f412c993ea74f64ec3962751313e1980d92074ca4","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-17T10:17:16Z","title_canon_sha256":"b3efce7e3030e59293df698065a15bb60510f6d3c385101269500421119326cf"},"schema_version":"1.0","source":{"id":"2604.16565","kind":"arxiv","version":3}},"canonical_sha256":"5e712ee34d3387ff364e6396455b47873837b572eb67bb677b205690ee447ad1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5e712ee34d3387ff364e6396455b47873837b572eb67bb677b205690ee447ad1","first_computed_at":"2026-05-28T02:04:47.980747Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-28T02:04:47.980747Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"m8184IUm+ylxD+xJ0iMdj97Q5zrBITRmmf3HtmOoLCF+GihCSn7NB05lpoP+YpwzJOK0UjM/SQHUQaf/UHqGCQ==","signature_status":"signed_v1","signed_at":"2026-05-28T02:04:47.981257Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.16565","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:43c188e5ef7d71766dd1b673f8b617379931b408691066a6e1c593211d232dce","sha256:0d16ec5f1ef963bc03f3f1849e6445a693f0080eef89778870d5f7ea0fb2b746"],"state_sha256":"bd70bc5ab4fe5d31b564c473eb5ea54dba1312d1ad027908d4311381bdc7b270"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HGwPsD1yLYkm0lYoUkfzXmRuFyX4UJy/PZkld3bZy9FoFPHVyid7Asufh62ruDRMl6HoPXdAH8Doyo/1dAaIAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-09T12:34:54.732365Z","bundle_sha256":"57fbc4e732da061741ac1cdd9b44f26a50a7513596e1925c7e2105db8dc3f492"}}