{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DX3BNS5IHQMWUSRZ6J4CCRS5TE","short_pith_number":"pith:DX3BNS5I","schema_version":"1.0","canonical_sha256":"1df616cba83c196a4a39f27821465d993f1c5784d49fb70e29590437162152fe","source":{"kind":"arxiv","id":"2605.18899","version":1},"attestation_state":"computed","paper":{"title":"Don't Let Bandit Feedback Pull Continual LLM-Recommender Updates Off Target","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chung Park, Hyeongjun Yun, Jaegul Choo, Taesan Kim","submitted_at":"2026-05-17T11:10:44Z","abstract_excerpt":"Generative LLM-based recommenders (LLM-Rec) require continual post-deployment updates, yet deployment logs provide only policy-shaped contextual bandit feedback: outcomes are observed solely for items exposed by a prior serving policy, inducing exposure bias and yielding partial, asymmetric signals consisting of relatively reliable positive responses and ambiguous no-responses. We propose an Anchored Bandit Policy Optimization (ABPO) framework for continual LLM-Rec updates that combines group-relative policy optimization (GRPO) with explicit treatment of exposure bias and feedback ambiguity. S"},"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":"2605.18899","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-17T11:10:44Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"2fc1b838bdcf9ce3db8424f3bfb3fd8e50005863e22ef286ae61b6e9dd0e2b60","abstract_canon_sha256":"4be6adf8cdc9a519c4708792d5a2189e9150b6249b7d66ec42e18bad28d010a5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:06:31.128856Z","signature_b64":"yrNur30hD/DmMXCwmZVI8uuFA12GqTptIoTi8D2ORireDlMc2nM8bmffAK5W0MvGQ1m1MmUHZkkEP50BfpvVBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1df616cba83c196a4a39f27821465d993f1c5784d49fb70e29590437162152fe","last_reissued_at":"2026-05-20T00:06:31.128071Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:06:31.128071Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Don't Let Bandit Feedback Pull Continual LLM-Recommender Updates Off Target","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chung Park, Hyeongjun Yun, Jaegul Choo, Taesan Kim","submitted_at":"2026-05-17T11:10:44Z","abstract_excerpt":"Generative LLM-based recommenders (LLM-Rec) require continual post-deployment updates, yet deployment logs provide only policy-shaped contextual bandit feedback: outcomes are observed solely for items exposed by a prior serving policy, inducing exposure bias and yielding partial, asymmetric signals consisting of relatively reliable positive responses and ambiguous no-responses. We propose an Anchored Bandit Policy Optimization (ABPO) framework for continual LLM-Rec updates that combines group-relative policy optimization (GRPO) with explicit treatment of exposure bias and feedback ambiguity. S"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.18899","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/2605.18899/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":"2605.18899","created_at":"2026-05-20T00:06:31.128201+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.18899v1","created_at":"2026-05-20T00:06:31.128201+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.18899","created_at":"2026-05-20T00:06:31.128201+00:00"},{"alias_kind":"pith_short_12","alias_value":"DX3BNS5IHQMW","created_at":"2026-05-20T00:06:31.128201+00:00"},{"alias_kind":"pith_short_16","alias_value":"DX3BNS5IHQMWUSRZ","created_at":"2026-05-20T00:06:31.128201+00:00"},{"alias_kind":"pith_short_8","alias_value":"DX3BNS5I","created_at":"2026-05-20T00:06:31.128201+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/DX3BNS5IHQMWUSRZ6J4CCRS5TE","json":"https://pith.science/pith/DX3BNS5IHQMWUSRZ6J4CCRS5TE.json","graph_json":"https://pith.science/api/pith-number/DX3BNS5IHQMWUSRZ6J4CCRS5TE/graph.json","events_json":"https://pith.science/api/pith-number/DX3BNS5IHQMWUSRZ6J4CCRS5TE/events.json","paper":"https://pith.science/paper/DX3BNS5I"},"agent_actions":{"view_html":"https://pith.science/pith/DX3BNS5IHQMWUSRZ6J4CCRS5TE","download_json":"https://pith.science/pith/DX3BNS5IHQMWUSRZ6J4CCRS5TE.json","view_paper":"https://pith.science/paper/DX3BNS5I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.18899&json=true","fetch_graph":"https://pith.science/api/pith-number/DX3BNS5IHQMWUSRZ6J4CCRS5TE/graph.json","fetch_events":"https://pith.science/api/pith-number/DX3BNS5IHQMWUSRZ6J4CCRS5TE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DX3BNS5IHQMWUSRZ6J4CCRS5TE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DX3BNS5IHQMWUSRZ6J4CCRS5TE/action/storage_attestation","attest_author":"https://pith.science/pith/DX3BNS5IHQMWUSRZ6J4CCRS5TE/action/author_attestation","sign_citation":"https://pith.science/pith/DX3BNS5IHQMWUSRZ6J4CCRS5TE/action/citation_signature","submit_replication":"https://pith.science/pith/DX3BNS5IHQMWUSRZ6J4CCRS5TE/action/replication_record"}},"created_at":"2026-05-20T00:06:31.128201+00:00","updated_at":"2026-05-20T00:06:31.128201+00:00"}