{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UVD2M6V345OVIN7MNMHSRAPL6R","short_pith_number":"pith:UVD2M6V3","schema_version":"1.0","canonical_sha256":"a547a67abbe75d5437ec6b0f2881ebf479bc276522d77007f44349189c3eb2c2","source":{"kind":"arxiv","id":"2606.23687","version":1},"attestation_state":"computed","paper":{"title":"Randomized YaRN Improves Length Generalization for Long-Context Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fangcong Yin, Greg Durrett, Manas Mehta","submitted_at":"2026-06-22T17:59:53Z","abstract_excerpt":"Large language models (LLMs) are typically pretrained on short sequences and then extended to work on longer sequences with additional training. However, such LLMs still struggle to further generalize to very long sequences. We propose Randomized YaRN, a training method that improves length generalization by combining YaRN-based positional extrapolation with randomized positional encoding and a length curriculum. During training on short context data, tokens are assigned YaRN positional encodings sampled from a larger position range, exposing the model to out-of-distribution positional represe"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.23687","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-06-22T17:59:53Z","cross_cats_sorted":[],"title_canon_sha256":"091f86da1bb421dd1683e38e53967be3148b4e43f9ccaf20e30a4eac7a922d66","abstract_canon_sha256":"34a507ffca7e5162382f0f3413c35ad473bd624f10f8091fca3993461e3e3459"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T03:14:34.806947Z","signature_b64":"Wb7mDyaBUcSgSJlmvmq3cE9zCtPSecCbuzIohRn8w6ebHOg3qMy7Sy3RLyUBQDVYqyH/DCnrpVW1XOeqWgJDCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a547a67abbe75d5437ec6b0f2881ebf479bc276522d77007f44349189c3eb2c2","last_reissued_at":"2026-06-23T03:14:34.806479Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T03:14:34.806479Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Randomized YaRN Improves Length Generalization for Long-Context Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fangcong Yin, Greg Durrett, Manas Mehta","submitted_at":"2026-06-22T17:59:53Z","abstract_excerpt":"Large language models (LLMs) are typically pretrained on short sequences and then extended to work on longer sequences with additional training. However, such LLMs still struggle to further generalize to very long sequences. We propose Randomized YaRN, a training method that improves length generalization by combining YaRN-based positional extrapolation with randomized positional encoding and a length curriculum. During training on short context data, tokens are assigned YaRN positional encodings sampled from a larger position range, exposing the model to out-of-distribution positional represe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.23687","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/2606.23687/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.23687","created_at":"2026-06-23T03:14:34.806540+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.23687v1","created_at":"2026-06-23T03:14:34.806540+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.23687","created_at":"2026-06-23T03:14:34.806540+00:00"},{"alias_kind":"pith_short_12","alias_value":"UVD2M6V345OV","created_at":"2026-06-23T03:14:34.806540+00:00"},{"alias_kind":"pith_short_16","alias_value":"UVD2M6V345OVIN7M","created_at":"2026-06-23T03:14:34.806540+00:00"},{"alias_kind":"pith_short_8","alias_value":"UVD2M6V3","created_at":"2026-06-23T03:14:34.806540+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R","json":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R.json","graph_json":"https://pith.science/api/pith-number/UVD2M6V345OVIN7MNMHSRAPL6R/graph.json","events_json":"https://pith.science/api/pith-number/UVD2M6V345OVIN7MNMHSRAPL6R/events.json","paper":"https://pith.science/paper/UVD2M6V3"},"agent_actions":{"view_html":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R","download_json":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R.json","view_paper":"https://pith.science/paper/UVD2M6V3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.23687&json=true","fetch_graph":"https://pith.science/api/pith-number/UVD2M6V345OVIN7MNMHSRAPL6R/graph.json","fetch_events":"https://pith.science/api/pith-number/UVD2M6V345OVIN7MNMHSRAPL6R/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R/action/storage_attestation","attest_author":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R/action/author_attestation","sign_citation":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R/action/citation_signature","submit_replication":"https://pith.science/pith/UVD2M6V345OVIN7MNMHSRAPL6R/action/replication_record"}},"created_at":"2026-06-23T03:14:34.806540+00:00","updated_at":"2026-06-23T03:14:34.806540+00:00"}