{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:Z3TEXB4A3OZ2YOUA2BDCZDSWMG","short_pith_number":"pith:Z3TEXB4A","canonical_record":{"source":{"id":"2502.13173","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-17T19:56:21Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"fec30a3953183e85f25c998b9a9264e499877523665b5b3e7b2c9cde1a0e23a8","abstract_canon_sha256":"d4782b976077e6048b5a3686c1a587c70f27f1ab9e2fddb9f61925bc0ecadb92"},"schema_version":"1.0"},"canonical_sha256":"cee64b8780dbb3ac3a80d0462c8e566193719fe915e6abe7de722171bf9ee003","source":{"kind":"arxiv","id":"2502.13173","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.13173","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"arxiv_version","alias_value":"2502.13173v1","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13173","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"pith_short_12","alias_value":"Z3TEXB4A3OZ2","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"pith_short_16","alias_value":"Z3TEXB4A3OZ2YOUA","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"pith_short_8","alias_value":"Z3TEXB4A","created_at":"2026-07-05T10:16:43Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:Z3TEXB4A3OZ2YOUA2BDCZDSWMG","target":"record","payload":{"canonical_record":{"source":{"id":"2502.13173","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-17T19:56:21Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"fec30a3953183e85f25c998b9a9264e499877523665b5b3e7b2c9cde1a0e23a8","abstract_canon_sha256":"d4782b976077e6048b5a3686c1a587c70f27f1ab9e2fddb9f61925bc0ecadb92"},"schema_version":"1.0"},"canonical_sha256":"cee64b8780dbb3ac3a80d0462c8e566193719fe915e6abe7de722171bf9ee003","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:16:43.430452Z","signature_b64":"JFusNVuNT2VHV7nwwTS9NUMIjgW3ZUfhDAiyVY9mL6QDW9r3I443B3j4MPgm0tcPjov+dptlZvdbGncHc73+Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cee64b8780dbb3ac3a80d0462c8e566193719fe915e6abe7de722171bf9ee003","last_reissued_at":"2026-07-05T10:16:43.429968Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:16:43.429968Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.13173","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:16:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lUkKAvCWujvUjnfiDDlXy7wGFni9E6IhBC28R/+twLX8nqKeVc/uZ90mc9VLOFOsUoiGY9icNeXc/oPD8KEwDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T11:14:02.510308Z"},"content_sha256":"6389767bb0746db48556c780532d487124a9434a990cfe0907b468ca924cb01c","schema_version":"1.0","event_id":"sha256:6389767bb0746db48556c780532d487124a9434a990cfe0907b468ca924cb01c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:Z3TEXB4A3OZ2YOUA2BDCZDSWMG","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Thinking Preference Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hongye Jin, Jingfeng Yang, Vipin Chaudhary, Wang Yang, Xiaotian Han","submitted_at":"2025-02-17T19:56:21Z","abstract_excerpt":"Supervised Fine-Tuning (SFT) has been a go-to and effective method for enhancing long chain-of-thought (CoT) reasoning in relatively small LLMs by fine-tuning them with long CoT responses from larger LLMs. To continually improve reasoning abilities, we can either collect new high-quality long CoT reasoning SFT data or repeatedly train on existing SFT datasets. However, acquiring new long CoT SFT data is costly and limited, while repeated training often results in a performance plateau or decline. To further boost the performance with the SFT data, we propose Thinking Preference Optimization (T"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13173","kind":"arxiv","version":1},"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/2502.13173/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":"2026-07-05T10:16:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kY1mu3yuxI96bMSsSitJWH+20jvK9kXiP1trGG7mfaNNjgfRnL4gAQdZsc3gjtBverUE+ImLInO2wqYDrN4VDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T11:14:02.510995Z"},"content_sha256":"c2641d793b0608432a10934d0a64fd5dd1d5a9c3486a7741d5ff0e2f40192e9b","schema_version":"1.0","event_id":"sha256:c2641d793b0608432a10934d0a64fd5dd1d5a9c3486a7741d5ff0e2f40192e9b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/Z3TEXB4A3OZ2YOUA2BDCZDSWMG/bundle.json","state_url":"https://pith.science/pith/Z3TEXB4A3OZ2YOUA2BDCZDSWMG/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/Z3TEXB4A3OZ2YOUA2BDCZDSWMG/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":"2026-08-08T11:14:02Z","links":{"resolver":"https://pith.science/pith/Z3TEXB4A3OZ2YOUA2BDCZDSWMG","bundle":"https://pith.science/pith/Z3TEXB4A3OZ2YOUA2BDCZDSWMG/bundle.json","state":"https://pith.science/pith/Z3TEXB4A3OZ2YOUA2BDCZDSWMG/state.json","well_known_bundle":"https://pith.science/.well-known/pith/Z3TEXB4A3OZ2YOUA2BDCZDSWMG/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:Z3TEXB4A3OZ2YOUA2BDCZDSWMG","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":"d4782b976077e6048b5a3686c1a587c70f27f1ab9e2fddb9f61925bc0ecadb92","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-17T19:56:21Z","title_canon_sha256":"fec30a3953183e85f25c998b9a9264e499877523665b5b3e7b2c9cde1a0e23a8"},"schema_version":"1.0","source":{"id":"2502.13173","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.13173","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"arxiv_version","alias_value":"2502.13173v1","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13173","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"pith_short_12","alias_value":"Z3TEXB4A3OZ2","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"pith_short_16","alias_value":"Z3TEXB4A3OZ2YOUA","created_at":"2026-07-05T10:16:43Z"},{"alias_kind":"pith_short_8","alias_value":"Z3TEXB4A","created_at":"2026-07-05T10:16:43Z"}],"graph_snapshots":[{"event_id":"sha256:c2641d793b0608432a10934d0a64fd5dd1d5a9c3486a7741d5ff0e2f40192e9b","target":"graph","created_at":"2026-07-05T10:16:43Z","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/2502.13173/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Supervised Fine-Tuning (SFT) has been a go-to and effective method for enhancing long chain-of-thought (CoT) reasoning in relatively small LLMs by fine-tuning them with long CoT responses from larger LLMs. To continually improve reasoning abilities, we can either collect new high-quality long CoT reasoning SFT data or repeatedly train on existing SFT datasets. However, acquiring new long CoT SFT data is costly and limited, while repeated training often results in a performance plateau or decline. To further boost the performance with the SFT data, we propose Thinking Preference Optimization (T","authors_text":"Hongye Jin, Jingfeng Yang, Vipin Chaudhary, Wang Yang, Xiaotian Han","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-17T19:56:21Z","title":"Thinking Preference Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13173","kind":"arxiv","version":1},"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":"sha256:6389767bb0746db48556c780532d487124a9434a990cfe0907b468ca924cb01c","target":"record","created_at":"2026-07-05T10:16:43Z","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":"d4782b976077e6048b5a3686c1a587c70f27f1ab9e2fddb9f61925bc0ecadb92","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-17T19:56:21Z","title_canon_sha256":"fec30a3953183e85f25c998b9a9264e499877523665b5b3e7b2c9cde1a0e23a8"},"schema_version":"1.0","source":{"id":"2502.13173","kind":"arxiv","version":1}},"canonical_sha256":"cee64b8780dbb3ac3a80d0462c8e566193719fe915e6abe7de722171bf9ee003","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"cee64b8780dbb3ac3a80d0462c8e566193719fe915e6abe7de722171bf9ee003","first_computed_at":"2026-07-05T10:16:43.429968Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:16:43.429968Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"JFusNVuNT2VHV7nwwTS9NUMIjgW3ZUfhDAiyVY9mL6QDW9r3I443B3j4MPgm0tcPjov+dptlZvdbGncHc73+Cg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:16:43.430452Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.13173","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6389767bb0746db48556c780532d487124a9434a990cfe0907b468ca924cb01c","sha256:c2641d793b0608432a10934d0a64fd5dd1d5a9c3486a7741d5ff0e2f40192e9b"],"state_sha256":"89f281bc3ad276ef93a08b8f17fbced71615dd46fd18ee6129f99cbf7ffe8415"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rnoHQ8IQd+5thX2g1Z+oDqHxMCJ7EbmGBSBh/LWyPqODv/6NLIYoWCvzqWaYCLkxu3/4SHwoFnBlBEui3D8VCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T11:14:02.516554Z","bundle_sha256":"f352f045cfde51c447c395db96a465fa3e3c840fc48c772287489d57f666ba42"}}