{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:DGK3WXKMHNN3ONONA5BRO454Q4","short_pith_number":"pith:DGK3WXKM","canonical_record":{"source":{"id":"2510.00526","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-01T05:17:47Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e487944db5c4d7d5f0102cc5117554b46d026d5c99a42e4fdb067f411bf8057a","abstract_canon_sha256":"056501924a14941e41fca4e62002e054ff432325126513adeaa1d499ca24baf5"},"schema_version":"1.0"},"canonical_sha256":"1995bb5d4c3b5bb735cd07431773bc870af6d3e0e8410ff66a5fee3f598457e6","source":{"kind":"arxiv","id":"2510.00526","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.00526","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"arxiv_version","alias_value":"2510.00526v3","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.00526","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"pith_short_12","alias_value":"DGK3WXKMHNN3","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"pith_short_16","alias_value":"DGK3WXKMHNN3ONON","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"pith_short_8","alias_value":"DGK3WXKM","created_at":"2026-05-25T02:01:07Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:DGK3WXKMHNN3ONONA5BRO454Q4","target":"record","payload":{"canonical_record":{"source":{"id":"2510.00526","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-01T05:17:47Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e487944db5c4d7d5f0102cc5117554b46d026d5c99a42e4fdb067f411bf8057a","abstract_canon_sha256":"056501924a14941e41fca4e62002e054ff432325126513adeaa1d499ca24baf5"},"schema_version":"1.0"},"canonical_sha256":"1995bb5d4c3b5bb735cd07431773bc870af6d3e0e8410ff66a5fee3f598457e6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:01:07.629665Z","signature_b64":"keYgPd+1hFFo7uyvBp7rX+FAoVgWeMBKU4O7xmWE+FFoiOHYoov7a6gJOH3r6A5BaR55lkwZWZHAxQ+uUlhVCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1995bb5d4c3b5bb735cd07431773bc870af6d3e0e8410ff66a5fee3f598457e6","last_reissued_at":"2026-05-25T02:01:07.628752Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:01:07.628752Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2510.00526","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":"2026-05-25T02:01:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/ZTo56LdfTRBQT7tsIfIXdt00fQLn3bDey1olQ8cS5yiR7d7x/l3vL46Mqywn43GqNJH4zZCX0bVYKedA+0CCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T04:06:57.884926Z"},"content_sha256":"4b815c69f37f878922bda1b3edecc6d256fb4549c8423a8ba21bc0ca0ca7e068","schema_version":"1.0","event_id":"sha256:4b815c69f37f878922bda1b3edecc6d256fb4549c8423a8ba21bc0ca0ca7e068"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:DGK3WXKMHNN3ONONA5BRO454Q4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Beyond Log Likelihood: Probability-Based Objectives for Supervised Fine-Tuning across the Model Capability Continuum","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Gaotang Li, Hanghang Tong, Heng Ji, Ruizhong Qiu, Xiusi Chen","submitted_at":"2025-10-01T05:17:47Z","abstract_excerpt":"Supervised fine-tuning (SFT) is the standard approach for post-training large language models (LLMs), yet it often shows limited generalization. We trace this limitation to its default training objective: negative log likelihood (NLL). While NLL is classically optimal when training from scratch, post-training operates in a different paradigm and could violate its optimality assumptions, where models already encode task-relevant priors and supervision can be long and noisy. In this work, we systematically study various probability-based objectives and characterize when and why different objecti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.00526","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/2510.00526/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":"2026-05-25T02:01:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0aeM+EqilO2wABII3qiRN74eWdwLOGdCscYynKA05Gj/vECS/v+RbXguHUfIRlv11pxx9+ss1MEdmbKCSdmMAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T04:06:57.885300Z"},"content_sha256":"329086f849bdcf7c98cc87c74dfd4e03035d6e9b4be8ea0299b357fe359aa730","schema_version":"1.0","event_id":"sha256:329086f849bdcf7c98cc87c74dfd4e03035d6e9b4be8ea0299b357fe359aa730"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DGK3WXKMHNN3ONONA5BRO454Q4/bundle.json","state_url":"https://pith.science/pith/DGK3WXKMHNN3ONONA5BRO454Q4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DGK3WXKMHNN3ONONA5BRO454Q4/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":"2026-06-02T04:06:57Z","links":{"resolver":"https://pith.science/pith/DGK3WXKMHNN3ONONA5BRO454Q4","bundle":"https://pith.science/pith/DGK3WXKMHNN3ONONA5BRO454Q4/bundle.json","state":"https://pith.science/pith/DGK3WXKMHNN3ONONA5BRO454Q4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DGK3WXKMHNN3ONONA5BRO454Q4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:DGK3WXKMHNN3ONONA5BRO454Q4","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":"056501924a14941e41fca4e62002e054ff432325126513adeaa1d499ca24baf5","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-01T05:17:47Z","title_canon_sha256":"e487944db5c4d7d5f0102cc5117554b46d026d5c99a42e4fdb067f411bf8057a"},"schema_version":"1.0","source":{"id":"2510.00526","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.00526","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"arxiv_version","alias_value":"2510.00526v3","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.00526","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"pith_short_12","alias_value":"DGK3WXKMHNN3","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"pith_short_16","alias_value":"DGK3WXKMHNN3ONON","created_at":"2026-05-25T02:01:07Z"},{"alias_kind":"pith_short_8","alias_value":"DGK3WXKM","created_at":"2026-05-25T02:01:07Z"}],"graph_snapshots":[{"event_id":"sha256:329086f849bdcf7c98cc87c74dfd4e03035d6e9b4be8ea0299b357fe359aa730","target":"graph","created_at":"2026-05-25T02:01:07Z","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/2510.00526/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Supervised fine-tuning (SFT) is the standard approach for post-training large language models (LLMs), yet it often shows limited generalization. We trace this limitation to its default training objective: negative log likelihood (NLL). While NLL is classically optimal when training from scratch, post-training operates in a different paradigm and could violate its optimality assumptions, where models already encode task-relevant priors and supervision can be long and noisy. In this work, we systematically study various probability-based objectives and characterize when and why different objecti","authors_text":"Gaotang Li, Hanghang Tong, Heng Ji, Ruizhong Qiu, Xiusi Chen","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-01T05:17:47Z","title":"Beyond Log Likelihood: Probability-Based Objectives for Supervised Fine-Tuning across the Model Capability Continuum"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.00526","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":"sha256:4b815c69f37f878922bda1b3edecc6d256fb4549c8423a8ba21bc0ca0ca7e068","target":"record","created_at":"2026-05-25T02:01:07Z","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":"056501924a14941e41fca4e62002e054ff432325126513adeaa1d499ca24baf5","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-01T05:17:47Z","title_canon_sha256":"e487944db5c4d7d5f0102cc5117554b46d026d5c99a42e4fdb067f411bf8057a"},"schema_version":"1.0","source":{"id":"2510.00526","kind":"arxiv","version":3}},"canonical_sha256":"1995bb5d4c3b5bb735cd07431773bc870af6d3e0e8410ff66a5fee3f598457e6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1995bb5d4c3b5bb735cd07431773bc870af6d3e0e8410ff66a5fee3f598457e6","first_computed_at":"2026-05-25T02:01:07.628752Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-25T02:01:07.628752Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"keYgPd+1hFFo7uyvBp7rX+FAoVgWeMBKU4O7xmWE+FFoiOHYoov7a6gJOH3r6A5BaR55lkwZWZHAxQ+uUlhVCw==","signature_status":"signed_v1","signed_at":"2026-05-25T02:01:07.629665Z","signed_message":"canonical_sha256_bytes"},"source_id":"2510.00526","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4b815c69f37f878922bda1b3edecc6d256fb4549c8423a8ba21bc0ca0ca7e068","sha256:329086f849bdcf7c98cc87c74dfd4e03035d6e9b4be8ea0299b357fe359aa730"],"state_sha256":"e61ca040b2c10a814cdf5112bc866a137865544a2e2708726081f9167cd86569"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"k4IP7owaDN3oiafY2BQRVj57ulrhQD9glqO92lyJS0nhJRzPIXwQ8USXiRRBV3ZqwGrJ85l8HwxTaHmzXmD+AA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-02T04:06:57.887299Z","bundle_sha256":"5f9ea9c13080811b998546028a2df61d5492b7a8424743fdf8c78fc1b8c2d18b"}}