{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:XSAGUHCWD3OASMV7BM5KXEGN44","short_pith_number":"pith:XSAGUHCW","canonical_record":{"source":{"id":"2510.13870","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-13T12:33:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"de91a36077a8ad4098cf168cca1de1369e557ab378e36d8b8822ce9778a7c090","abstract_canon_sha256":"1c98e3ec3fe9d30ab499b2767b85c70b7e8234a13f8ddf5de192d8d3f2b32055"},"schema_version":"1.0"},"canonical_sha256":"bc806a1c561edc0932bf0b3aab90cde724bc43dbd27b1cce4fe49ab8042af8d5","source":{"kind":"arxiv","id":"2510.13870","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.13870","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"arxiv_version","alias_value":"2510.13870v3","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.13870","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"pith_short_12","alias_value":"XSAGUHCWD3OA","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"pith_short_16","alias_value":"XSAGUHCWD3OASMV7","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"pith_short_8","alias_value":"XSAGUHCW","created_at":"2026-05-20T00:04:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:XSAGUHCWD3OASMV7BM5KXEGN44","target":"record","payload":{"canonical_record":{"source":{"id":"2510.13870","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-13T12:33:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"de91a36077a8ad4098cf168cca1de1369e557ab378e36d8b8822ce9778a7c090","abstract_canon_sha256":"1c98e3ec3fe9d30ab499b2767b85c70b7e8234a13f8ddf5de192d8d3f2b32055"},"schema_version":"1.0"},"canonical_sha256":"bc806a1c561edc0932bf0b3aab90cde724bc43dbd27b1cce4fe49ab8042af8d5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:04:14.783559Z","signature_b64":"0ItRkfdk5Pil5W5oqOt6o52UC1ayvOAyV5YS15fiP3GrGAdUmrRp9Hpe/vAy11CmY31UOj9vC/Fd0xftFjpQCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bc806a1c561edc0932bf0b3aab90cde724bc43dbd27b1cce4fe49ab8042af8d5","last_reissued_at":"2026-05-20T00:04:14.782681Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:04:14.782681Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2510.13870","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":"2026-05-20T00:04:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Ze+OWKjA28nd8KHLUsKZKB+P7kPp7IbeTZpU+nbcSvU9mYXHQucPNy3LMYOlS6VNPfG6pNCu6Cqvsze9IlC+BA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T13:26:38.302264Z"},"content_sha256":"08677b4546ad437a8410922a5333a3a52ad10f1af9ad79c437963f519f20d1ab","schema_version":"1.0","event_id":"sha256:08677b4546ad437a8410922a5333a3a52ad10f1af9ad79c437963f519f20d1ab"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:XSAGUHCWD3OASMV7BM5KXEGN44","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Unlocking the Potential of Diffusion Language Models through Template Infilling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Template Infilling aligns structural anchors across the full response to guide diffusion language models before filling details.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Junhoo Lee, Nojun Kwak, Seungyeon Kim","submitted_at":"2025-10-13T12:33:41Z","abstract_excerpt":"Diffusion Language Models (DLMs) have emerged as a promising alternative to Autoregressive Language Models, yet their inference strategies remain limited to prefix-based prompting inherited from the autoregressive paradigm. In this paper, we propose Template Infilling (TI), a tailored conditioning methodology for DLMs. Unlike conventional prefix prompting, TI flexibly aligns structural anchors across the entire target response space, establishing a global blueprint before filling in the masked segments. We demonstrate the effectiveness of our approach on diverse benchmarks, including mathemati"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Template Infilling flexibly aligns structural anchors across the entire target response space, establishing a global blueprint before filling in the masked segments, and achieves consistent improvements of 9.40% over the baseline on mathematical reasoning, code generation, and trip planning.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That providing structural anchors in a template will reliably guide the diffusion denoising process to respect global constraints without degrading local coherence or requiring additional training, as the abstract presents this as the key mechanism enabling the reported gains.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Template Infilling improves diffusion language models by aligning structural anchors across the entire response space for global constraints before infilling, yielding 9.4% gains on math, code, and planning benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Template Infilling aligns structural anchors across the full response to guide diffusion language models before filling details.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f2b7d0ad90759b92da333022e75f940d3d03ccf05ebcc843347223fb841eb4c3"},"source":{"id":"2510.13870","kind":"arxiv","version":3},"verdict":{"id":"6788558e-8e18-4c42-8901-b0bab6a6a9fd","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-18T07:47:01.408102Z","strongest_claim":"Template Infilling flexibly aligns structural anchors across the entire target response space, establishing a global blueprint before filling in the masked segments, and achieves consistent improvements of 9.40% over the baseline on mathematical reasoning, code generation, and trip planning.","one_line_summary":"Template Infilling improves diffusion language models by aligning structural anchors across the entire response space for global constraints before infilling, yielding 9.4% gains on math, code, and planning benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That providing structural anchors in a template will reliably guide the diffusion denoising process to respect global constraints without degrading local coherence or requiring additional training, as the abstract presents this as the key mechanism enabling the reported gains.","pith_extraction_headline":"Template Infilling aligns structural anchors across the full response to guide diffusion language models before filling details."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2510.13870/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":2,"snapshot_sha256":"61a806f156bd37da94b177e4209bfa023c1dab3677ee71417f7750190d19fcad"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"6788558e-8e18-4c42-8901-b0bab6a6a9fd"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:04:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qScr4BIfKazoo62EzXYsinsQAZAHAaOlyzFQ/QUXZ/XrDSRhlnSn/M4/oKUdZCACOwI5r/ZWYQQrAdqCDQ4IBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T13:26:38.303124Z"},"content_sha256":"cc979cdfdb6f86ef1e505795ab93512250bcac2e5d737f1ca3cad92b6811db5e","schema_version":"1.0","event_id":"sha256:cc979cdfdb6f86ef1e505795ab93512250bcac2e5d737f1ca3cad92b6811db5e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/XSAGUHCWD3OASMV7BM5KXEGN44/bundle.json","state_url":"https://pith.science/pith/XSAGUHCWD3OASMV7BM5KXEGN44/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/XSAGUHCWD3OASMV7BM5KXEGN44/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":"2026-05-25T13:26:38Z","links":{"resolver":"https://pith.science/pith/XSAGUHCWD3OASMV7BM5KXEGN44","bundle":"https://pith.science/pith/XSAGUHCWD3OASMV7BM5KXEGN44/bundle.json","state":"https://pith.science/pith/XSAGUHCWD3OASMV7BM5KXEGN44/state.json","well_known_bundle":"https://pith.science/.well-known/pith/XSAGUHCWD3OASMV7BM5KXEGN44/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:XSAGUHCWD3OASMV7BM5KXEGN44","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":"1c98e3ec3fe9d30ab499b2767b85c70b7e8234a13f8ddf5de192d8d3f2b32055","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-13T12:33:41Z","title_canon_sha256":"de91a36077a8ad4098cf168cca1de1369e557ab378e36d8b8822ce9778a7c090"},"schema_version":"1.0","source":{"id":"2510.13870","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.13870","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"arxiv_version","alias_value":"2510.13870v3","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.13870","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"pith_short_12","alias_value":"XSAGUHCWD3OA","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"pith_short_16","alias_value":"XSAGUHCWD3OASMV7","created_at":"2026-05-20T00:04:14Z"},{"alias_kind":"pith_short_8","alias_value":"XSAGUHCW","created_at":"2026-05-20T00:04:14Z"}],"graph_snapshots":[{"event_id":"sha256:cc979cdfdb6f86ef1e505795ab93512250bcac2e5d737f1ca3cad92b6811db5e","target":"graph","created_at":"2026-05-20T00:04:14Z","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":"Template Infilling flexibly aligns structural anchors across the entire target response space, establishing a global blueprint before filling in the masked segments, and achieves consistent improvements of 9.40% over the baseline on mathematical reasoning, code generation, and trip planning."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That providing structural anchors in a template will reliably guide the diffusion denoising process to respect global constraints without degrading local coherence or requiring additional training, as the abstract presents this as the key mechanism enabling the reported gains."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Template Infilling improves diffusion language models by aligning structural anchors across the entire response space for global constraints before infilling, yielding 9.4% gains on math, code, and planning benchmarks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Template Infilling aligns structural anchors across the full response to guide diffusion language models before filling details."}],"snapshot_sha256":"f2b7d0ad90759b92da333022e75f940d3d03ccf05ebcc843347223fb841eb4c3"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"61a806f156bd37da94b177e4209bfa023c1dab3677ee71417f7750190d19fcad"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2510.13870/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Diffusion Language Models (DLMs) have emerged as a promising alternative to Autoregressive Language Models, yet their inference strategies remain limited to prefix-based prompting inherited from the autoregressive paradigm. In this paper, we propose Template Infilling (TI), a tailored conditioning methodology for DLMs. Unlike conventional prefix prompting, TI flexibly aligns structural anchors across the entire target response space, establishing a global blueprint before filling in the masked segments. We demonstrate the effectiveness of our approach on diverse benchmarks, including mathemati","authors_text":"Junhoo Lee, Nojun Kwak, Seungyeon Kim","cross_cats":["cs.AI"],"headline":"Template Infilling aligns structural anchors across the full response to guide diffusion language models before filling details.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-13T12:33:41Z","title":"Unlocking the Potential of Diffusion Language Models through Template Infilling"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.13870","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-18T07:47:01.408102Z","id":"6788558e-8e18-4c42-8901-b0bab6a6a9fd","model_set":{"reader":"grok-4.3"},"one_line_summary":"Template Infilling improves diffusion language models by aligning structural anchors across the entire response space for global constraints before infilling, yielding 9.4% gains on math, code, and planning benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Template Infilling aligns structural anchors across the full response to guide diffusion language models before filling details.","strongest_claim":"Template Infilling flexibly aligns structural anchors across the entire target response space, establishing a global blueprint before filling in the masked segments, and achieves consistent improvements of 9.40% over the baseline on mathematical reasoning, code generation, and trip planning.","weakest_assumption":"That providing structural anchors in a template will reliably guide the diffusion denoising process to respect global constraints without degrading local coherence or requiring additional training, as the abstract presents this as the key mechanism enabling the reported gains."}},"verdict_id":"6788558e-8e18-4c42-8901-b0bab6a6a9fd"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:08677b4546ad437a8410922a5333a3a52ad10f1af9ad79c437963f519f20d1ab","target":"record","created_at":"2026-05-20T00:04:14Z","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":"1c98e3ec3fe9d30ab499b2767b85c70b7e8234a13f8ddf5de192d8d3f2b32055","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-13T12:33:41Z","title_canon_sha256":"de91a36077a8ad4098cf168cca1de1369e557ab378e36d8b8822ce9778a7c090"},"schema_version":"1.0","source":{"id":"2510.13870","kind":"arxiv","version":3}},"canonical_sha256":"bc806a1c561edc0932bf0b3aab90cde724bc43dbd27b1cce4fe49ab8042af8d5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"bc806a1c561edc0932bf0b3aab90cde724bc43dbd27b1cce4fe49ab8042af8d5","first_computed_at":"2026-05-20T00:04:14.782681Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:04:14.782681Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"0ItRkfdk5Pil5W5oqOt6o52UC1ayvOAyV5YS15fiP3GrGAdUmrRp9Hpe/vAy11CmY31UOj9vC/Fd0xftFjpQCQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:04:14.783559Z","signed_message":"canonical_sha256_bytes"},"source_id":"2510.13870","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:08677b4546ad437a8410922a5333a3a52ad10f1af9ad79c437963f519f20d1ab","sha256:cc979cdfdb6f86ef1e505795ab93512250bcac2e5d737f1ca3cad92b6811db5e"],"state_sha256":"c00ca25aa2ad88166efcc7d852ccb12935b8c57718809535568cb916e7323550"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jn8ToqSQAaAKrLPxDj/VULXtvL02vSyv2b1gaiGklho5ctSfBpnl4uSifwhkuavJ50OQE+XWVwXyZFHbh+fMCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-25T13:26:38.306848Z","bundle_sha256":"62670c2b8eb5d9e18b9817c86190c980d32c2ac1f6c8851fb89e08b8a53d67e5"}}