{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:SH6ECHJNA2FNUC7IE6TWEF32PT","short_pith_number":"pith:SH6ECHJN","canonical_record":{"source":{"id":"2508.12837","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-18T11:24:30Z","cross_cats_sorted":[],"title_canon_sha256":"ab151cf787220feb8e6f29a232dc44f43cddce0842167746ce27ef0d54a5001b","abstract_canon_sha256":"bb53462a8d9d4f1e6f327101984a1d45afdb5c9df0e2ab1b1b0d49512fe6e7a2"},"schema_version":"1.0"},"canonical_sha256":"91fc411d2d068ada0be827a762177a7cfe365c3d8a03b1cb8c2655b179b15d53","source":{"kind":"arxiv","id":"2508.12837","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.12837","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"arxiv_version","alias_value":"2508.12837v2","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.12837","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"pith_short_12","alias_value":"SH6ECHJNA2FN","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"pith_short_16","alias_value":"SH6ECHJNA2FNUC7I","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"pith_short_8","alias_value":"SH6ECHJN","created_at":"2026-07-05T11:55:43Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:SH6ECHJNA2FNUC7IE6TWEF32PT","target":"record","payload":{"canonical_record":{"source":{"id":"2508.12837","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-18T11:24:30Z","cross_cats_sorted":[],"title_canon_sha256":"ab151cf787220feb8e6f29a232dc44f43cddce0842167746ce27ef0d54a5001b","abstract_canon_sha256":"bb53462a8d9d4f1e6f327101984a1d45afdb5c9df0e2ab1b1b0d49512fe6e7a2"},"schema_version":"1.0"},"canonical_sha256":"91fc411d2d068ada0be827a762177a7cfe365c3d8a03b1cb8c2655b179b15d53","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:55:43.007188Z","signature_b64":"0ub4gfh2UMVQS9OXea4VJXkIg/02GdndnfNADr3NdBSu0B7i4BYUiwu/sVrD60FrPXfn6JZaZsVCaMvrYjLlCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"91fc411d2d068ada0be827a762177a7cfe365c3d8a03b1cb8c2655b179b15d53","last_reissued_at":"2026-07-05T11:55:43.006767Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:55:43.006767Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.12837","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:55:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MLdoEr1n1GRM8PTmYqofwy+y5vYcsGmUnd9csQ14mBMkx7elkNvkFZsiCS9B5Z17yDPKwPDXSEy9eHzImIw0BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T12:34:54.011638Z"},"content_sha256":"b85a2ad230ff6210befe123b2e986760b59a72b71da6af3c2d7df80749ed380c","schema_version":"1.0","event_id":"sha256:b85a2ad230ff6210befe123b2e986760b59a72b71da6af3c2d7df80749ed380c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:SH6ECHJNA2FNUC7IE6TWEF32PT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning In-context n-grams with Transformers: Sub-n-grams Are Near-stationary Points","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aditya Varre, Gizem Y\\\"uce, Nicolas Flammarion","submitted_at":"2025-08-18T11:24:30Z","abstract_excerpt":"Motivated by empirical observations of prolonged plateaus and stage-wise progression during training, we investigate the loss landscape of transformer models trained on in-context next-token prediction tasks. In particular, we focus on learning in-context $n$-gram language models under cross-entropy loss, and establish a sufficient condition for parameter configurations to be stationary points. We then construct a set of parameter configurations for a simplified transformer model that represent $k$-gram estimators (for $k \\leq n$), and show that the gradient of the population loss at these sol"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.12837","kind":"arxiv","version":2},"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/2508.12837/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":"2026-07-05T11:55:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nEE4HCEuDSkKKsamWFgaLcPKtkurAEEtQ+JTDTAc8L+dcBaI4DfVxJQ1cJM3rbfUPStpH+oQJ692RAZCXMarAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T12:34:54.012540Z"},"content_sha256":"6012588664888934d3b66dfbef6ce1edec88fa72c0a6ef983c0351583c024b32","schema_version":"1.0","event_id":"sha256:6012588664888934d3b66dfbef6ce1edec88fa72c0a6ef983c0351583c024b32"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SH6ECHJNA2FNUC7IE6TWEF32PT/bundle.json","state_url":"https://pith.science/pith/SH6ECHJNA2FNUC7IE6TWEF32PT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SH6ECHJNA2FNUC7IE6TWEF32PT/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":"2026-08-04T12:34:54Z","links":{"resolver":"https://pith.science/pith/SH6ECHJNA2FNUC7IE6TWEF32PT","bundle":"https://pith.science/pith/SH6ECHJNA2FNUC7IE6TWEF32PT/bundle.json","state":"https://pith.science/pith/SH6ECHJNA2FNUC7IE6TWEF32PT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SH6ECHJNA2FNUC7IE6TWEF32PT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:SH6ECHJNA2FNUC7IE6TWEF32PT","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":"bb53462a8d9d4f1e6f327101984a1d45afdb5c9df0e2ab1b1b0d49512fe6e7a2","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-18T11:24:30Z","title_canon_sha256":"ab151cf787220feb8e6f29a232dc44f43cddce0842167746ce27ef0d54a5001b"},"schema_version":"1.0","source":{"id":"2508.12837","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.12837","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"arxiv_version","alias_value":"2508.12837v2","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.12837","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"pith_short_12","alias_value":"SH6ECHJNA2FN","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"pith_short_16","alias_value":"SH6ECHJNA2FNUC7I","created_at":"2026-07-05T11:55:43Z"},{"alias_kind":"pith_short_8","alias_value":"SH6ECHJN","created_at":"2026-07-05T11:55:43Z"}],"graph_snapshots":[{"event_id":"sha256:6012588664888934d3b66dfbef6ce1edec88fa72c0a6ef983c0351583c024b32","target":"graph","created_at":"2026-07-05T11:55:43Z","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/2508.12837/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Motivated by empirical observations of prolonged plateaus and stage-wise progression during training, we investigate the loss landscape of transformer models trained on in-context next-token prediction tasks. In particular, we focus on learning in-context $n$-gram language models under cross-entropy loss, and establish a sufficient condition for parameter configurations to be stationary points. We then construct a set of parameter configurations for a simplified transformer model that represent $k$-gram estimators (for $k \\leq n$), and show that the gradient of the population loss at these sol","authors_text":"Aditya Varre, Gizem Y\\\"uce, Nicolas Flammarion","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-18T11:24:30Z","title":"Learning In-context n-grams with Transformers: Sub-n-grams Are Near-stationary Points"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.12837","kind":"arxiv","version":2},"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:b85a2ad230ff6210befe123b2e986760b59a72b71da6af3c2d7df80749ed380c","target":"record","created_at":"2026-07-05T11:55:43Z","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":"bb53462a8d9d4f1e6f327101984a1d45afdb5c9df0e2ab1b1b0d49512fe6e7a2","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-18T11:24:30Z","title_canon_sha256":"ab151cf787220feb8e6f29a232dc44f43cddce0842167746ce27ef0d54a5001b"},"schema_version":"1.0","source":{"id":"2508.12837","kind":"arxiv","version":2}},"canonical_sha256":"91fc411d2d068ada0be827a762177a7cfe365c3d8a03b1cb8c2655b179b15d53","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"91fc411d2d068ada0be827a762177a7cfe365c3d8a03b1cb8c2655b179b15d53","first_computed_at":"2026-07-05T11:55:43.006767Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:55:43.006767Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"0ub4gfh2UMVQS9OXea4VJXkIg/02GdndnfNADr3NdBSu0B7i4BYUiwu/sVrD60FrPXfn6JZaZsVCaMvrYjLlCw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:55:43.007188Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.12837","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b85a2ad230ff6210befe123b2e986760b59a72b71da6af3c2d7df80749ed380c","sha256:6012588664888934d3b66dfbef6ce1edec88fa72c0a6ef983c0351583c024b32"],"state_sha256":"1388ffad29024e37897c7e908bb49b4357a5cedfa99890cd0a3b95d506edc984"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"N9DUKVeedh7dsgWvcij0DdNHWpOzHx9Yo5KUiPHYV/m/N+zg3sTO7e+2r5LH4bL7JzVGSnEhSLxE/T0s1f2GDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T12:34:54.019639Z","bundle_sha256":"608269bd9b87b5a2b56c167f03202d0eeea4cd899eb4dad1dfbaa3f5ef24fda6"}}