{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:VZMYBS7BCRY33EYVESV5MXNDWQ","short_pith_number":"pith:VZMYBS7B","schema_version":"1.0","canonical_sha256":"ae5980cbe11471bd931524abd65da3b439605201ca902db32e7eeacd62d1804d","source":{"kind":"arxiv","id":"2606.06902","version":1},"attestation_state":"computed","paper":{"title":"TALAN: Task-Aligned Latent Adaptation Networks for Targeted Post-Training of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chengkai Zhang, Junpu Wang, Qin Huang, Sagar Chordia, Yang Wang, Zeyi Tao, Ziteng Liu","submitted_at":"2026-06-05T04:34:50Z","abstract_excerpt":"Targeted post-training aims to improve reasoning, math, and code without degrading strengths. Low-rank adapters are efficient but task-global; activation interventions are input-aware but often require separate probes, vectors, or inference-time steering. We introduce TALAN (Task-Aligned Latent Adaptation Networks), a sequence-conditioned latent side path inserted into a transformer's residual stream and co-trained with a low-rank adapter in one SFT loop. TALAN compresses the active sequence into latent memory, remixes it into token-level perturbations, and writes them back through a controlle"},"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":"2606.06902","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-05T04:34:50Z","cross_cats_sorted":[],"title_canon_sha256":"30128f9b9fdafbd76d90203bb1cef357afec2440cf1e1c1ed66ace8d3bd188a0","abstract_canon_sha256":"db29228ffe5f89926455280d8dcb038f3db5a0f73a37b4e1e3c67975941b9ec4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-08T01:04:34.806658Z","signature_b64":"duwArHeCM73Mu4afWcAzfiiBaDlhPudlNnzOCkmzCaQKxYnqJQypKNPHgHHRzYWe/cbNPihZRNElI8ImCAVGAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ae5980cbe11471bd931524abd65da3b439605201ca902db32e7eeacd62d1804d","last_reissued_at":"2026-06-08T01:04:34.805794Z","signature_status":"signed_v1","first_computed_at":"2026-06-08T01:04:34.805794Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TALAN: Task-Aligned Latent Adaptation Networks for Targeted Post-Training of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chengkai Zhang, Junpu Wang, Qin Huang, Sagar Chordia, Yang Wang, Zeyi Tao, Ziteng Liu","submitted_at":"2026-06-05T04:34:50Z","abstract_excerpt":"Targeted post-training aims to improve reasoning, math, and code without degrading strengths. Low-rank adapters are efficient but task-global; activation interventions are input-aware but often require separate probes, vectors, or inference-time steering. We introduce TALAN (Task-Aligned Latent Adaptation Networks), a sequence-conditioned latent side path inserted into a transformer's residual stream and co-trained with a low-rank adapter in one SFT loop. TALAN compresses the active sequence into latent memory, remixes it into token-level perturbations, and writes them back through a controlle"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.06902","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/2606.06902/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":"2606.06902","created_at":"2026-06-08T01:04:34.805927+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.06902v1","created_at":"2026-06-08T01:04:34.805927+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.06902","created_at":"2026-06-08T01:04:34.805927+00:00"},{"alias_kind":"pith_short_12","alias_value":"VZMYBS7BCRY3","created_at":"2026-06-08T01:04:34.805927+00:00"},{"alias_kind":"pith_short_16","alias_value":"VZMYBS7BCRY33EYV","created_at":"2026-06-08T01:04:34.805927+00:00"},{"alias_kind":"pith_short_8","alias_value":"VZMYBS7B","created_at":"2026-06-08T01:04:34.805927+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/VZMYBS7BCRY33EYVESV5MXNDWQ","json":"https://pith.science/pith/VZMYBS7BCRY33EYVESV5MXNDWQ.json","graph_json":"https://pith.science/api/pith-number/VZMYBS7BCRY33EYVESV5MXNDWQ/graph.json","events_json":"https://pith.science/api/pith-number/VZMYBS7BCRY33EYVESV5MXNDWQ/events.json","paper":"https://pith.science/paper/VZMYBS7B"},"agent_actions":{"view_html":"https://pith.science/pith/VZMYBS7BCRY33EYVESV5MXNDWQ","download_json":"https://pith.science/pith/VZMYBS7BCRY33EYVESV5MXNDWQ.json","view_paper":"https://pith.science/paper/VZMYBS7B","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.06902&json=true","fetch_graph":"https://pith.science/api/pith-number/VZMYBS7BCRY33EYVESV5MXNDWQ/graph.json","fetch_events":"https://pith.science/api/pith-number/VZMYBS7BCRY33EYVESV5MXNDWQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VZMYBS7BCRY33EYVESV5MXNDWQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VZMYBS7BCRY33EYVESV5MXNDWQ/action/storage_attestation","attest_author":"https://pith.science/pith/VZMYBS7BCRY33EYVESV5MXNDWQ/action/author_attestation","sign_citation":"https://pith.science/pith/VZMYBS7BCRY33EYVESV5MXNDWQ/action/citation_signature","submit_replication":"https://pith.science/pith/VZMYBS7BCRY33EYVESV5MXNDWQ/action/replication_record"}},"created_at":"2026-06-08T01:04:34.805927+00:00","updated_at":"2026-06-08T01:04:34.805927+00:00"}