{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:N7MDMRPOJGIFASEQ4B7JWOA4SJ","short_pith_number":"pith:N7MDMRPO","canonical_record":{"source":{"id":"2501.04945","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-09T03:34:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"540fd83f5a37bfc45c68ccf83bb5a3e6c7146ad1cffc51e86c3c9d47a5532a33","abstract_canon_sha256":"12cb559e4af8bea02a758c8cce075ae2acddb52ee93ede387147662bc3f3d25f"},"schema_version":"1.0"},"canonical_sha256":"6fd83645ee4990504890e07e9b381c9268b03467b9f402bc8600696317c55036","source":{"kind":"arxiv","id":"2501.04945","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.04945","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"arxiv_version","alias_value":"2501.04945v4","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.04945","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"pith_short_12","alias_value":"N7MDMRPOJGIF","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"pith_short_16","alias_value":"N7MDMRPOJGIFASEQ","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"pith_short_8","alias_value":"N7MDMRPO","created_at":"2026-07-05T11:13:18Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:N7MDMRPOJGIFASEQ4B7JWOA4SJ","target":"record","payload":{"canonical_record":{"source":{"id":"2501.04945","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-09T03:34:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"540fd83f5a37bfc45c68ccf83bb5a3e6c7146ad1cffc51e86c3c9d47a5532a33","abstract_canon_sha256":"12cb559e4af8bea02a758c8cce075ae2acddb52ee93ede387147662bc3f3d25f"},"schema_version":"1.0"},"canonical_sha256":"6fd83645ee4990504890e07e9b381c9268b03467b9f402bc8600696317c55036","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:13:18.484753Z","signature_b64":"zKFsenpzv/rtKZMR1MLXqvftvra90W3W+JQjSgsCk1tnYx8lcREQw0KSV20wFHevluqsUO8wBO6fcqa6J+ALCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6fd83645ee4990504890e07e9b381c9268b03467b9f402bc8600696317c55036","last_reissued_at":"2026-07-05T11:13:18.484264Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:13:18.484264Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2501.04945","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:13:18Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/1evjCamO3ETlst5WusXRivKhXsNHxkEtbNIg/r33yJKnymSjr5jFmI/SR3HDjkicXpq7nEyW3FiXCaLMtKiCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T14:06:36.593171Z"},"content_sha256":"3768dc8903167a5778b9d64f3dbe5bc80041192b1209d65e30d1697ff9ecf473","schema_version":"1.0","event_id":"sha256:3768dc8903167a5778b9d64f3dbe5bc80041192b1209d65e30d1697ff9ecf473"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:N7MDMRPOJGIFASEQ4B7JWOA4SJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Step-by-Step Mastery: Enhancing Soft Constraint Following Ability of Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Fei Yu, Jiaqing Liang, Jie Zeng, Qianyu He, Qingyu Ren, Weikang Zhou, Yanghua Xiao, Zeye Sun","submitted_at":"2025-01-09T03:34:07Z","abstract_excerpt":"It is crucial for large language models (LLMs) to follow instructions that involve multiple constraints. However, it is an unexplored area to enhance LLMs' ability to follow soft constraints. To bridge the gap, we initially design a pipeline to construct datasets with high-quality outputs automatically. Additionally, to fully utilize the positive and negative samples generated during the data construction process, we choose Direct Preference Optimization (DPO) as the training method. Furthermore, taking into account the difficulty of soft constraints indicated by the number of constraints, we "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.04945","kind":"arxiv","version":4},"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/2501.04945/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":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:13:18Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7b43LhFUBgv2EgXavQw8je7sswXpvfk8SYWd8j0BPQOvwQvZil5RU0AaxMahwBF1ncbN8TWQcqMmVVR/lD+BDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T14:06:36.593557Z"},"content_sha256":"3ff161540816ebc409bc53fe1017ec8f7649afa0758b0fb7f6857bf8ce59a84c","schema_version":"1.0","event_id":"sha256:3ff161540816ebc409bc53fe1017ec8f7649afa0758b0fb7f6857bf8ce59a84c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/N7MDMRPOJGIFASEQ4B7JWOA4SJ/bundle.json","state_url":"https://pith.science/pith/N7MDMRPOJGIFASEQ4B7JWOA4SJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/N7MDMRPOJGIFASEQ4B7JWOA4SJ/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":"2026-07-07T14:06:36Z","links":{"resolver":"https://pith.science/pith/N7MDMRPOJGIFASEQ4B7JWOA4SJ","bundle":"https://pith.science/pith/N7MDMRPOJGIFASEQ4B7JWOA4SJ/bundle.json","state":"https://pith.science/pith/N7MDMRPOJGIFASEQ4B7JWOA4SJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/N7MDMRPOJGIFASEQ4B7JWOA4SJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:N7MDMRPOJGIFASEQ4B7JWOA4SJ","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":"12cb559e4af8bea02a758c8cce075ae2acddb52ee93ede387147662bc3f3d25f","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-09T03:34:07Z","title_canon_sha256":"540fd83f5a37bfc45c68ccf83bb5a3e6c7146ad1cffc51e86c3c9d47a5532a33"},"schema_version":"1.0","source":{"id":"2501.04945","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.04945","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"arxiv_version","alias_value":"2501.04945v4","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.04945","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"pith_short_12","alias_value":"N7MDMRPOJGIF","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"pith_short_16","alias_value":"N7MDMRPOJGIFASEQ","created_at":"2026-07-05T11:13:18Z"},{"alias_kind":"pith_short_8","alias_value":"N7MDMRPO","created_at":"2026-07-05T11:13:18Z"}],"graph_snapshots":[{"event_id":"sha256:3ff161540816ebc409bc53fe1017ec8f7649afa0758b0fb7f6857bf8ce59a84c","target":"graph","created_at":"2026-07-05T11:13:18Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2501.04945/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"It is crucial for large language models (LLMs) to follow instructions that involve multiple constraints. However, it is an unexplored area to enhance LLMs' ability to follow soft constraints. To bridge the gap, we initially design a pipeline to construct datasets with high-quality outputs automatically. Additionally, to fully utilize the positive and negative samples generated during the data construction process, we choose Direct Preference Optimization (DPO) as the training method. Furthermore, taking into account the difficulty of soft constraints indicated by the number of constraints, we ","authors_text":"Fei Yu, Jiaqing Liang, Jie Zeng, Qianyu He, Qingyu Ren, Weikang Zhou, Yanghua Xiao, Zeye Sun","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-09T03:34:07Z","title":"Step-by-Step Mastery: Enhancing Soft Constraint Following Ability of Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.04945","kind":"arxiv","version":4},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:3768dc8903167a5778b9d64f3dbe5bc80041192b1209d65e30d1697ff9ecf473","target":"record","created_at":"2026-07-05T11:13:18Z","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":"12cb559e4af8bea02a758c8cce075ae2acddb52ee93ede387147662bc3f3d25f","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-09T03:34:07Z","title_canon_sha256":"540fd83f5a37bfc45c68ccf83bb5a3e6c7146ad1cffc51e86c3c9d47a5532a33"},"schema_version":"1.0","source":{"id":"2501.04945","kind":"arxiv","version":4}},"canonical_sha256":"6fd83645ee4990504890e07e9b381c9268b03467b9f402bc8600696317c55036","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6fd83645ee4990504890e07e9b381c9268b03467b9f402bc8600696317c55036","first_computed_at":"2026-07-05T11:13:18.484264Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:13:18.484264Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zKFsenpzv/rtKZMR1MLXqvftvra90W3W+JQjSgsCk1tnYx8lcREQw0KSV20wFHevluqsUO8wBO6fcqa6J+ALCw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:13:18.484753Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.04945","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3768dc8903167a5778b9d64f3dbe5bc80041192b1209d65e30d1697ff9ecf473","sha256:3ff161540816ebc409bc53fe1017ec8f7649afa0758b0fb7f6857bf8ce59a84c"],"state_sha256":"5c7160c1e8c0025b45c70fb825add5ea64203af4f6acd47e4a5ab069decf55e9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3ljIzZe2JJfWKts2EwaYeRfgE8BBubNqF4tEi0bjTau2hdYlNVO2vEF6C0YwAoJoL8zwEzjAxOeNVzw1uaObAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T14:06:36.595582Z","bundle_sha256":"9a2359108fb6eeab85d5dab47293c878f2480cf7faddc73355d5047dac5d47a9"}}