{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:FJGVIXSB3PL5E3YHH2X4RRITCJ","short_pith_number":"pith:FJGVIXSB","canonical_record":{"source":{"id":"2405.14734","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-23T16:01:46Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"2b61fa617e45adee7eab764326074219c6ed83d896e6bcaffe2920dda6bf1352","abstract_canon_sha256":"689d90003b54e394ceb86d19bb542feeee437354eb6b87a0ac6b03fcc7cfc858"},"schema_version":"1.0"},"canonical_sha256":"2a4d545e41dbd7d26f073eafc8c51312681431b31fd5f1f11b21dc417bf06748","source":{"kind":"arxiv","id":"2405.14734","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.14734","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"arxiv_version","alias_value":"2405.14734v3","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.14734","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"pith_short_12","alias_value":"FJGVIXSB3PL5","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"pith_short_16","alias_value":"FJGVIXSB3PL5E3YH","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"pith_short_8","alias_value":"FJGVIXSB","created_at":"2026-07-05T09:29:49Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:FJGVIXSB3PL5E3YHH2X4RRITCJ","target":"record","payload":{"canonical_record":{"source":{"id":"2405.14734","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-23T16:01:46Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"2b61fa617e45adee7eab764326074219c6ed83d896e6bcaffe2920dda6bf1352","abstract_canon_sha256":"689d90003b54e394ceb86d19bb542feeee437354eb6b87a0ac6b03fcc7cfc858"},"schema_version":"1.0"},"canonical_sha256":"2a4d545e41dbd7d26f073eafc8c51312681431b31fd5f1f11b21dc417bf06748","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:29:49.176087Z","signature_b64":"R7jV73QfyDH2zSkyXCbNcByg0JQay33VqINNzsEzIw7fapH9FFmCnrkcDY7FL7psfEHZPW6YNsgLeDJWKc4/CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2a4d545e41dbd7d26f073eafc8c51312681431b31fd5f1f11b21dc417bf06748","last_reissued_at":"2026-07-05T09:29:49.175626Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:29:49.175626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.14734","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":"2026-07-05T09:29:49Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Ntnj1VfTUOuBC20C6GFGxcTuIAPgcwBUmTMht5oEF6nayImpjEpy8jfWn5IEJ/p8u094TfC0epI0nHiao/7pAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-12T07:00:22.134584Z"},"content_sha256":"e9b83d1f77bf43c5dad8de6b02de6ec05986c745866a25519b1c9cac99c59b5a","schema_version":"1.0","event_id":"sha256:e9b83d1f77bf43c5dad8de6b02de6ec05986c745866a25519b1c9cac99c59b5a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:FJGVIXSB3PL5E3YHH2X4RRITCJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SimPO: Simple Preference Optimization with a Reference-Free Reward","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Danqi Chen, Mengzhou Xia, Yu Meng","submitted_at":"2024-05-23T16:01:46Z","abstract_excerpt":"Direct Preference Optimization (DPO) is a widely used offline preference optimization algorithm that reparameterizes reward functions in reinforcement learning from human feedback (RLHF) to enhance simplicity and training stability. In this work, we propose SimPO, a simpler yet more effective approach. The effectiveness of SimPO is attributed to a key design: using the average log probability of a sequence as the implicit reward. This reward formulation better aligns with model generation and eliminates the need for a reference model, making it more compute and memory efficient. Additionally, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.14734","kind":"arxiv","version":3},"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/2405.14734/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":"2026-07-05T09:29:49Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kahHhuNVv5VcYjJ3ctNjuLlWgVmfOc1WjnD09wfsUhNDal8VJv5YCC9dbivYJ8Nu8SwLYe+bUnZ/Oxk5xQ3OAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-12T07:00:22.134950Z"},"content_sha256":"0d1ac3bfd3bdeaa01f3b17686ff038a55225354c04fe19703c5dda9c8dd12185","schema_version":"1.0","event_id":"sha256:0d1ac3bfd3bdeaa01f3b17686ff038a55225354c04fe19703c5dda9c8dd12185"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FJGVIXSB3PL5E3YHH2X4RRITCJ/bundle.json","state_url":"https://pith.science/pith/FJGVIXSB3PL5E3YHH2X4RRITCJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FJGVIXSB3PL5E3YHH2X4RRITCJ/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":"2026-07-12T07:00:22Z","links":{"resolver":"https://pith.science/pith/FJGVIXSB3PL5E3YHH2X4RRITCJ","bundle":"https://pith.science/pith/FJGVIXSB3PL5E3YHH2X4RRITCJ/bundle.json","state":"https://pith.science/pith/FJGVIXSB3PL5E3YHH2X4RRITCJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FJGVIXSB3PL5E3YHH2X4RRITCJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:FJGVIXSB3PL5E3YHH2X4RRITCJ","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":"689d90003b54e394ceb86d19bb542feeee437354eb6b87a0ac6b03fcc7cfc858","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-23T16:01:46Z","title_canon_sha256":"2b61fa617e45adee7eab764326074219c6ed83d896e6bcaffe2920dda6bf1352"},"schema_version":"1.0","source":{"id":"2405.14734","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.14734","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"arxiv_version","alias_value":"2405.14734v3","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.14734","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"pith_short_12","alias_value":"FJGVIXSB3PL5","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"pith_short_16","alias_value":"FJGVIXSB3PL5E3YH","created_at":"2026-07-05T09:29:49Z"},{"alias_kind":"pith_short_8","alias_value":"FJGVIXSB","created_at":"2026-07-05T09:29:49Z"}],"graph_snapshots":[{"event_id":"sha256:0d1ac3bfd3bdeaa01f3b17686ff038a55225354c04fe19703c5dda9c8dd12185","target":"graph","created_at":"2026-07-05T09:29:49Z","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/2405.14734/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Direct Preference Optimization (DPO) is a widely used offline preference optimization algorithm that reparameterizes reward functions in reinforcement learning from human feedback (RLHF) to enhance simplicity and training stability. In this work, we propose SimPO, a simpler yet more effective approach. The effectiveness of SimPO is attributed to a key design: using the average log probability of a sequence as the implicit reward. This reward formulation better aligns with model generation and eliminates the need for a reference model, making it more compute and memory efficient. Additionally, ","authors_text":"Danqi Chen, Mengzhou Xia, Yu Meng","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-23T16:01:46Z","title":"SimPO: Simple Preference Optimization with a Reference-Free Reward"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.14734","kind":"arxiv","version":3},"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:e9b83d1f77bf43c5dad8de6b02de6ec05986c745866a25519b1c9cac99c59b5a","target":"record","created_at":"2026-07-05T09:29:49Z","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":"689d90003b54e394ceb86d19bb542feeee437354eb6b87a0ac6b03fcc7cfc858","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-23T16:01:46Z","title_canon_sha256":"2b61fa617e45adee7eab764326074219c6ed83d896e6bcaffe2920dda6bf1352"},"schema_version":"1.0","source":{"id":"2405.14734","kind":"arxiv","version":3}},"canonical_sha256":"2a4d545e41dbd7d26f073eafc8c51312681431b31fd5f1f11b21dc417bf06748","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2a4d545e41dbd7d26f073eafc8c51312681431b31fd5f1f11b21dc417bf06748","first_computed_at":"2026-07-05T09:29:49.175626Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:29:49.175626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"R7jV73QfyDH2zSkyXCbNcByg0JQay33VqINNzsEzIw7fapH9FFmCnrkcDY7FL7psfEHZPW6YNsgLeDJWKc4/CA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:29:49.176087Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.14734","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e9b83d1f77bf43c5dad8de6b02de6ec05986c745866a25519b1c9cac99c59b5a","sha256:0d1ac3bfd3bdeaa01f3b17686ff038a55225354c04fe19703c5dda9c8dd12185"],"state_sha256":"25f65602fe52f3e3ba7fac0be352176696e7ae83c6ff6b2f96ced2de1f58b16e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HNSce7gm27X8pat2n3aNe6dTzDDz7s+2e3td9OHqdj36ok0aP0aJ8YoT2emmSsgUDHfN3V8854aKWlRD6TcVAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-12T07:00:22.137480Z","bundle_sha256":"3d333f7bde388e1804fe9983a70a272f058905827d6efe9d28a5c460951e09ea"}}