{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:5QNPMYK555MVNHKN4C7BSHMAGW","short_pith_number":"pith:5QNPMYK5","canonical_record":{"source":{"id":"2505.20144","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-26T15:45:56Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f46e7d14c0ffdc20b30336eeb02621e2e3f13158096f8fde6b64dd58e4812e9e","abstract_canon_sha256":"740761cc27cb650afdf324acdffb60eb698da63506cce06dbb57c9db87575d7d"},"schema_version":"1.0"},"canonical_sha256":"ec1af6615def59569d4de0be191d8035be0d4c7677c5a8848ae7bbb770995498","source":{"kind":"arxiv","id":"2505.20144","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.20144","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"arxiv_version","alias_value":"2505.20144v1","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.20144","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"pith_short_12","alias_value":"5QNPMYK555MV","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"pith_short_16","alias_value":"5QNPMYK555MVNHKN","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"pith_short_8","alias_value":"5QNPMYK5","created_at":"2026-07-05T11:09:53Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:5QNPMYK555MVNHKN4C7BSHMAGW","target":"record","payload":{"canonical_record":{"source":{"id":"2505.20144","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-26T15:45:56Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f46e7d14c0ffdc20b30336eeb02621e2e3f13158096f8fde6b64dd58e4812e9e","abstract_canon_sha256":"740761cc27cb650afdf324acdffb60eb698da63506cce06dbb57c9db87575d7d"},"schema_version":"1.0"},"canonical_sha256":"ec1af6615def59569d4de0be191d8035be0d4c7677c5a8848ae7bbb770995498","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:09:53.183772Z","signature_b64":"dwZmUAWCoYaFvi1CEGx6Gk7qqLy418lJTMBQx7W5SIYPNL0M//LlRyGfv2EKmoc4EVwD6OxIUrvAwX8hb0ieBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ec1af6615def59569d4de0be191d8035be0d4c7677c5a8848ae7bbb770995498","last_reissued_at":"2026-07-05T11:09:53.183276Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:09:53.183276Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.20144","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":"2026-07-05T11:09:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4pt16jU9ToPXdxRd1GPvmdxUhjYyy2nG+C54jjyzTCnd2VGRS12FZkdUqpIhDMzZrC2zVG4PE971r5iDq262AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T05:11:42.108007Z"},"content_sha256":"9df4e5e7d2e703fecf3740a9e3f2112a50cc23313906d81ceb4461193ae2ecec","schema_version":"1.0","event_id":"sha256:9df4e5e7d2e703fecf3740a9e3f2112a50cc23313906d81ceb4461193ae2ecec"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:5QNPMYK555MVNHKN4C7BSHMAGW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SeMe: Training-Free Language Model Merging via Semantic Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Aldeida Aleti, Chunyang Chen, Hongyu Zhang, Jian Gu","submitted_at":"2025-05-26T15:45:56Z","abstract_excerpt":"Despite the remarkable capabilities of Language Models (LMs) across diverse tasks, no single model consistently outperforms others, necessitating efficient methods to combine their strengths without expensive retraining. Existing model merging techniques, such as parameter averaging and task-guided fusion, often rely on data-dependent computations or fail to preserve internal knowledge, limiting their robustness and scalability. We introduce SeMe (Semantic-based Merging), a novel, data-free, and training-free approach that leverages latent semantic alignment to merge LMs at a fine-grained, lay"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.20144","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/2505.20144/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":"2026-07-05T11:09:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"U5aC2xpy7FHISCNpTxxgH9sUqVVtfPBaq812fNDY/g69V2XKEVf7rA7WFngzLkh/Oi6E871SGpoCTWX3nKH9AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T05:11:42.108940Z"},"content_sha256":"7be7b84dba61233ab649990055ee4c5655e6d54af9111ce35a4397091b77dff1","schema_version":"1.0","event_id":"sha256:7be7b84dba61233ab649990055ee4c5655e6d54af9111ce35a4397091b77dff1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5QNPMYK555MVNHKN4C7BSHMAGW/bundle.json","state_url":"https://pith.science/pith/5QNPMYK555MVNHKN4C7BSHMAGW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5QNPMYK555MVNHKN4C7BSHMAGW/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":"2026-08-06T05:11:42Z","links":{"resolver":"https://pith.science/pith/5QNPMYK555MVNHKN4C7BSHMAGW","bundle":"https://pith.science/pith/5QNPMYK555MVNHKN4C7BSHMAGW/bundle.json","state":"https://pith.science/pith/5QNPMYK555MVNHKN4C7BSHMAGW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5QNPMYK555MVNHKN4C7BSHMAGW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:5QNPMYK555MVNHKN4C7BSHMAGW","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":"740761cc27cb650afdf324acdffb60eb698da63506cce06dbb57c9db87575d7d","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-26T15:45:56Z","title_canon_sha256":"f46e7d14c0ffdc20b30336eeb02621e2e3f13158096f8fde6b64dd58e4812e9e"},"schema_version":"1.0","source":{"id":"2505.20144","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.20144","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"arxiv_version","alias_value":"2505.20144v1","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.20144","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"pith_short_12","alias_value":"5QNPMYK555MV","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"pith_short_16","alias_value":"5QNPMYK555MVNHKN","created_at":"2026-07-05T11:09:53Z"},{"alias_kind":"pith_short_8","alias_value":"5QNPMYK5","created_at":"2026-07-05T11:09:53Z"}],"graph_snapshots":[{"event_id":"sha256:7be7b84dba61233ab649990055ee4c5655e6d54af9111ce35a4397091b77dff1","target":"graph","created_at":"2026-07-05T11:09:53Z","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/2505.20144/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite the remarkable capabilities of Language Models (LMs) across diverse tasks, no single model consistently outperforms others, necessitating efficient methods to combine their strengths without expensive retraining. Existing model merging techniques, such as parameter averaging and task-guided fusion, often rely on data-dependent computations or fail to preserve internal knowledge, limiting their robustness and scalability. We introduce SeMe (Semantic-based Merging), a novel, data-free, and training-free approach that leverages latent semantic alignment to merge LMs at a fine-grained, lay","authors_text":"Aldeida Aleti, Chunyang Chen, Hongyu Zhang, Jian Gu","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-26T15:45:56Z","title":"SeMe: Training-Free Language Model Merging via Semantic Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.20144","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":"sha256:9df4e5e7d2e703fecf3740a9e3f2112a50cc23313906d81ceb4461193ae2ecec","target":"record","created_at":"2026-07-05T11:09:53Z","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":"740761cc27cb650afdf324acdffb60eb698da63506cce06dbb57c9db87575d7d","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-26T15:45:56Z","title_canon_sha256":"f46e7d14c0ffdc20b30336eeb02621e2e3f13158096f8fde6b64dd58e4812e9e"},"schema_version":"1.0","source":{"id":"2505.20144","kind":"arxiv","version":1}},"canonical_sha256":"ec1af6615def59569d4de0be191d8035be0d4c7677c5a8848ae7bbb770995498","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ec1af6615def59569d4de0be191d8035be0d4c7677c5a8848ae7bbb770995498","first_computed_at":"2026-07-05T11:09:53.183276Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:09:53.183276Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"dwZmUAWCoYaFvi1CEGx6Gk7qqLy418lJTMBQx7W5SIYPNL0M//LlRyGfv2EKmoc4EVwD6OxIUrvAwX8hb0ieBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:09:53.183772Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.20144","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9df4e5e7d2e703fecf3740a9e3f2112a50cc23313906d81ceb4461193ae2ecec","sha256:7be7b84dba61233ab649990055ee4c5655e6d54af9111ce35a4397091b77dff1"],"state_sha256":"77afa8080dd3d9adc3a55067ba8e1f68003042f0711cf1afd5e8845620434c23"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/kQevnFGeG/9Nq9HqQoWP3YrHsPdb7aoGMBTrCGNSu6rSD6RcJ5vH61ukZ6A0e4ki4uhy8PgmtJ0yZ3xOjcmDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T05:11:42.116160Z","bundle_sha256":"4c02d39c5fcb4428b53a2dff5707d52e3b14c0336d1917c64bfdd8702a8b2f45"}}