{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:5ZWFXW5RGGA7WHGTPPC5KGJMIM","short_pith_number":"pith:5ZWFXW5R","canonical_record":{"source":{"id":"2212.09849","kind":"arxiv","version":6},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-19T20:46:43Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"724852c2ab7b1e64ceddd8aa840fd497ca46c29d1795b0e839cae7d739c1f8cc","abstract_canon_sha256":"83f9f90b031f259d0e84c54dab0f6240544dec47aabf938d8d582e7352218445"},"schema_version":"1.0"},"canonical_sha256":"ee6c5bdbb13181fb1cd37bc5d5192c433624cc3987703324a096607fcd187bef","source":{"kind":"arxiv","id":"2212.09849","version":6},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2212.09849","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"arxiv_version","alias_value":"2212.09849v6","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.09849","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"pith_short_12","alias_value":"5ZWFXW5RGGA7","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"pith_short_16","alias_value":"5ZWFXW5RGGA7WHGT","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"pith_short_8","alias_value":"5ZWFXW5R","created_at":"2026-07-05T11:06:59Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:5ZWFXW5RGGA7WHGTPPC5KGJMIM","target":"record","payload":{"canonical_record":{"source":{"id":"2212.09849","kind":"arxiv","version":6},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-19T20:46:43Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"724852c2ab7b1e64ceddd8aa840fd497ca46c29d1795b0e839cae7d739c1f8cc","abstract_canon_sha256":"83f9f90b031f259d0e84c54dab0f6240544dec47aabf938d8d582e7352218445"},"schema_version":"1.0"},"canonical_sha256":"ee6c5bdbb13181fb1cd37bc5d5192c433624cc3987703324a096607fcd187bef","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:06:59.465184Z","signature_b64":"HnShSx+7NDH5r7gCrGmMkHXKjb9NBHpG81nySXHBjuG30FMIgWAQf2U0fTu0qQFDWztyV2QcN3juIW5QXfx6Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ee6c5bdbb13181fb1cd37bc5d5192c433624cc3987703324a096607fcd187bef","last_reissued_at":"2026-07-05T11:06:59.464729Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:06:59.464729Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2212.09849","source_version":6,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:06:59Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uam0T77dorMrNI2MEnVNGLOXBWtQ68dqiL7ZwRzeUaQxtIX5fGfts3LAxhUBjTpZCb/d6SSBLK6nwUBFrwbUCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T11:08:12.041387Z"},"content_sha256":"4ab2d8222f15842e2ed393b7c38b77048cfafef3f6bbdd1d72dbca23ca380a8b","schema_version":"1.0","event_id":"sha256:4ab2d8222f15842e2ed393b7c38b77048cfafef3f6bbdd1d72dbca23ca380a8b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:5ZWFXW5RGGA7WHGTPPC5KGJMIM","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Dataless Knowledge Fusion by Merging Weights of Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Daniel Preotiuc-Pietro, Pengxiang Cheng, Xiang Ren, Xisen Jin","submitted_at":"2022-12-19T20:46:43Z","abstract_excerpt":"Fine-tuning pre-trained language models has become the prevalent paradigm for building downstream NLP models. Oftentimes fine-tuned models are readily available but their training data is not, due to data privacy or intellectual property concerns. This creates a barrier to fusing knowledge across individual models to yield a better single model. In this paper, we study the problem of merging individual models built on different training data sets to obtain a single model that performs well both across all data set domains and can generalize on out-of-domain data. We propose a dataless knowledg"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.09849","kind":"arxiv","version":6},"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/2212.09849/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":"2026-07-05T11:06:59Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7LWP1UJES7wgeMQga9EHTBA2GCHwIdEzyFtsHBBXT097HLul3byY8M9atsu3Gd0x8eEUQ4EivVNakTToPCsPCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T11:08:12.051423Z"},"content_sha256":"18adf13795e9a00e3a4845b8e1ed90a01c6f6133ba409e12e0a5218d1dc4ee2a","schema_version":"1.0","event_id":"sha256:18adf13795e9a00e3a4845b8e1ed90a01c6f6133ba409e12e0a5218d1dc4ee2a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5ZWFXW5RGGA7WHGTPPC5KGJMIM/bundle.json","state_url":"https://pith.science/pith/5ZWFXW5RGGA7WHGTPPC5KGJMIM/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5ZWFXW5RGGA7WHGTPPC5KGJMIM/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":"2026-08-05T11:08:12Z","links":{"resolver":"https://pith.science/pith/5ZWFXW5RGGA7WHGTPPC5KGJMIM","bundle":"https://pith.science/pith/5ZWFXW5RGGA7WHGTPPC5KGJMIM/bundle.json","state":"https://pith.science/pith/5ZWFXW5RGGA7WHGTPPC5KGJMIM/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5ZWFXW5RGGA7WHGTPPC5KGJMIM/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:5ZWFXW5RGGA7WHGTPPC5KGJMIM","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":"83f9f90b031f259d0e84c54dab0f6240544dec47aabf938d8d582e7352218445","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-19T20:46:43Z","title_canon_sha256":"724852c2ab7b1e64ceddd8aa840fd497ca46c29d1795b0e839cae7d739c1f8cc"},"schema_version":"1.0","source":{"id":"2212.09849","kind":"arxiv","version":6}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2212.09849","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"arxiv_version","alias_value":"2212.09849v6","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.09849","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"pith_short_12","alias_value":"5ZWFXW5RGGA7","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"pith_short_16","alias_value":"5ZWFXW5RGGA7WHGT","created_at":"2026-07-05T11:06:59Z"},{"alias_kind":"pith_short_8","alias_value":"5ZWFXW5R","created_at":"2026-07-05T11:06:59Z"}],"graph_snapshots":[{"event_id":"sha256:18adf13795e9a00e3a4845b8e1ed90a01c6f6133ba409e12e0a5218d1dc4ee2a","target":"graph","created_at":"2026-07-05T11:06:59Z","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/2212.09849/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Fine-tuning pre-trained language models has become the prevalent paradigm for building downstream NLP models. Oftentimes fine-tuned models are readily available but their training data is not, due to data privacy or intellectual property concerns. This creates a barrier to fusing knowledge across individual models to yield a better single model. In this paper, we study the problem of merging individual models built on different training data sets to obtain a single model that performs well both across all data set domains and can generalize on out-of-domain data. We propose a dataless knowledg","authors_text":"Daniel Preotiuc-Pietro, Pengxiang Cheng, Xiang Ren, Xisen Jin","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-19T20:46:43Z","title":"Dataless Knowledge Fusion by Merging Weights of Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.09849","kind":"arxiv","version":6},"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":"sha256:4ab2d8222f15842e2ed393b7c38b77048cfafef3f6bbdd1d72dbca23ca380a8b","target":"record","created_at":"2026-07-05T11:06:59Z","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":"83f9f90b031f259d0e84c54dab0f6240544dec47aabf938d8d582e7352218445","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-19T20:46:43Z","title_canon_sha256":"724852c2ab7b1e64ceddd8aa840fd497ca46c29d1795b0e839cae7d739c1f8cc"},"schema_version":"1.0","source":{"id":"2212.09849","kind":"arxiv","version":6}},"canonical_sha256":"ee6c5bdbb13181fb1cd37bc5d5192c433624cc3987703324a096607fcd187bef","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ee6c5bdbb13181fb1cd37bc5d5192c433624cc3987703324a096607fcd187bef","first_computed_at":"2026-07-05T11:06:59.464729Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:06:59.464729Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HnShSx+7NDH5r7gCrGmMkHXKjb9NBHpG81nySXHBjuG30FMIgWAQf2U0fTu0qQFDWztyV2QcN3juIW5QXfx6Dw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:06:59.465184Z","signed_message":"canonical_sha256_bytes"},"source_id":"2212.09849","source_kind":"arxiv","source_version":6}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4ab2d8222f15842e2ed393b7c38b77048cfafef3f6bbdd1d72dbca23ca380a8b","sha256:18adf13795e9a00e3a4845b8e1ed90a01c6f6133ba409e12e0a5218d1dc4ee2a"],"state_sha256":"f45364586f30e979821e1e0ff71e21291644ca29233fa67e8ae31a8b9f74bbc3"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3iu0mc9ccxj5QTfEAlkRQLwf3oXRB2YJmkY0vIY6kM4KPM4Xu2irlF0U3b7MmL+isBS/5UgEH3i9H1A7MmWADw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T11:08:12.076013Z","bundle_sha256":"90ead0be10602436e04443a2c846ef35ad2b2bd55c815057c713593ff10e71e5"}}