{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:6RDEEJZ74HDZ2BNOJP3GYJEU6E","short_pith_number":"pith:6RDEEJZ7","canonical_record":{"source":{"id":"2508.04149","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T07:24:14Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"d7754273c29b1f20e16c82fca7121d7f49decdfa93fc2839e6ef324946704c6c","abstract_canon_sha256":"3effde57a54ab1ce8aacd8278f43a92e2e77aefd94836498c3d2e06d23fb380c"},"schema_version":"1.0"},"canonical_sha256":"f44642273fe1c79d05ae4bf66c2494f1089f0c0ba926d5bb968652f704dc2f37","source":{"kind":"arxiv","id":"2508.04149","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.04149","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"arxiv_version","alias_value":"2508.04149v2","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.04149","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"pith_short_12","alias_value":"6RDEEJZ74HDZ","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"pith_short_16","alias_value":"6RDEEJZ74HDZ2BNO","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"pith_short_8","alias_value":"6RDEEJZ7","created_at":"2026-05-20T00:02:52Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:6RDEEJZ74HDZ2BNOJP3GYJEU6E","target":"record","payload":{"canonical_record":{"source":{"id":"2508.04149","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T07:24:14Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"d7754273c29b1f20e16c82fca7121d7f49decdfa93fc2839e6ef324946704c6c","abstract_canon_sha256":"3effde57a54ab1ce8aacd8278f43a92e2e77aefd94836498c3d2e06d23fb380c"},"schema_version":"1.0"},"canonical_sha256":"f44642273fe1c79d05ae4bf66c2494f1089f0c0ba926d5bb968652f704dc2f37","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:02:52.603656Z","signature_b64":"kYmVmt2/SkLbWBWRded6bC9q/hXAPnq9ZQavO6kTTiu17RQpBwHQbDG2w+0IVvMXCrM27CBObhnf5V4exEf3DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f44642273fe1c79d05ae4bf66c2494f1089f0c0ba926d5bb968652f704dc2f37","last_reissued_at":"2026-05-20T00:02:52.602696Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:02:52.602696Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.04149","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:02:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4urgHBG7+6erG9qLZ24is7NfdBiaiH3uotvyqnE1qrDfbk8FCYTppYnnhzUGYsVuhGT+0cliq5Yaso7IKux7Ag==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-28T01:26:52.278115Z"},"content_sha256":"b151048071875c5ce645e705ca900f05d47dcf0241ae5d7186283bac3acb7f0f","schema_version":"1.0","event_id":"sha256:b151048071875c5ce645e705ca900f05d47dcf0241ae5d7186283bac3acb7f0f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:6RDEEJZ74HDZ2BNOJP3GYJEU6E","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Difficulty-Based Preference Data Selection by DPO Implicit Reward Gap","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Rongwu Xu, Xuan Qi, Zhijing Jin","submitted_at":"2025-08-06T07:24:14Z","abstract_excerpt":"Aligning large language models (LLMs) with human preferences is a critical challenge in AI research. While methods like Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) are widely used, they often rely on large, costly preference datasets. The current work lacks methods for high-quality data selection specifically for preference data. In this work, we introduce a novel difficulty-based data selection strategy for preference datasets, grounded in the DPO implicit reward mechanism. By selecting preference data examples with smaller DPO implicit reward ga"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.04149","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/2508.04149/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":"2026-05-20T00:02:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ATPX6dQUXVPpxmGbft5VtP1/SC+L+QuVkybTQJhToI9Hag7zA3fuSZy0FVUP8aE0D4ce8kY4uqSCV3Qq68csCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-28T01:26:52.278523Z"},"content_sha256":"c446985c9bef4df054cc9d6b460aa44d83172ed1ab357bdf32778a97b7b0dea1","schema_version":"1.0","event_id":"sha256:c446985c9bef4df054cc9d6b460aa44d83172ed1ab357bdf32778a97b7b0dea1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6RDEEJZ74HDZ2BNOJP3GYJEU6E/bundle.json","state_url":"https://pith.science/pith/6RDEEJZ74HDZ2BNOJP3GYJEU6E/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6RDEEJZ74HDZ2BNOJP3GYJEU6E/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":"2026-05-28T01:26:52Z","links":{"resolver":"https://pith.science/pith/6RDEEJZ74HDZ2BNOJP3GYJEU6E","bundle":"https://pith.science/pith/6RDEEJZ74HDZ2BNOJP3GYJEU6E/bundle.json","state":"https://pith.science/pith/6RDEEJZ74HDZ2BNOJP3GYJEU6E/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6RDEEJZ74HDZ2BNOJP3GYJEU6E/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:6RDEEJZ74HDZ2BNOJP3GYJEU6E","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":"3effde57a54ab1ce8aacd8278f43a92e2e77aefd94836498c3d2e06d23fb380c","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T07:24:14Z","title_canon_sha256":"d7754273c29b1f20e16c82fca7121d7f49decdfa93fc2839e6ef324946704c6c"},"schema_version":"1.0","source":{"id":"2508.04149","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.04149","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"arxiv_version","alias_value":"2508.04149v2","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.04149","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"pith_short_12","alias_value":"6RDEEJZ74HDZ","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"pith_short_16","alias_value":"6RDEEJZ74HDZ2BNO","created_at":"2026-05-20T00:02:52Z"},{"alias_kind":"pith_short_8","alias_value":"6RDEEJZ7","created_at":"2026-05-20T00:02:52Z"}],"graph_snapshots":[{"event_id":"sha256:c446985c9bef4df054cc9d6b460aa44d83172ed1ab357bdf32778a97b7b0dea1","target":"graph","created_at":"2026-05-20T00:02:52Z","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/2508.04149/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Aligning large language models (LLMs) with human preferences is a critical challenge in AI research. While methods like Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) are widely used, they often rely on large, costly preference datasets. The current work lacks methods for high-quality data selection specifically for preference data. In this work, we introduce a novel difficulty-based data selection strategy for preference datasets, grounded in the DPO implicit reward mechanism. By selecting preference data examples with smaller DPO implicit reward ga","authors_text":"Rongwu Xu, Xuan Qi, Zhijing Jin","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T07:24:14Z","title":"Difficulty-Based Preference Data Selection by DPO Implicit Reward Gap"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.04149","kind":"arxiv","version":2},"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:b151048071875c5ce645e705ca900f05d47dcf0241ae5d7186283bac3acb7f0f","target":"record","created_at":"2026-05-20T00:02:52Z","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":"3effde57a54ab1ce8aacd8278f43a92e2e77aefd94836498c3d2e06d23fb380c","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T07:24:14Z","title_canon_sha256":"d7754273c29b1f20e16c82fca7121d7f49decdfa93fc2839e6ef324946704c6c"},"schema_version":"1.0","source":{"id":"2508.04149","kind":"arxiv","version":2}},"canonical_sha256":"f44642273fe1c79d05ae4bf66c2494f1089f0c0ba926d5bb968652f704dc2f37","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f44642273fe1c79d05ae4bf66c2494f1089f0c0ba926d5bb968652f704dc2f37","first_computed_at":"2026-05-20T00:02:52.602696Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:02:52.602696Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"kYmVmt2/SkLbWBWRded6bC9q/hXAPnq9ZQavO6kTTiu17RQpBwHQbDG2w+0IVvMXCrM27CBObhnf5V4exEf3DA==","signature_status":"signed_v1","signed_at":"2026-05-20T00:02:52.603656Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.04149","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b151048071875c5ce645e705ca900f05d47dcf0241ae5d7186283bac3acb7f0f","sha256:c446985c9bef4df054cc9d6b460aa44d83172ed1ab357bdf32778a97b7b0dea1"],"state_sha256":"dc781746484f766ec7d2b29f6a973ea24553d4391b7999a308901e6b5e249333"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"imoURfq4s/qiV0cxuDvJkfXlab8m7SQlnrKQrtXDDobOiAEPirEuH+9NgOJUV5MWwhitfjv+qJtX0XIyTQF0CA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-28T01:26:52.280637Z","bundle_sha256":"57d8451b0686ba2810d14831632d9c45f3d5f741388280fe8b0eaaff7437a52b"}}