{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:N4VTAIPGAWHWCEEBFQTLZ2JRRC","short_pith_number":"pith:N4VTAIPG","schema_version":"1.0","canonical_sha256":"6f2b3021e6058f6110812c26bce93188b27cc6f4627e41613b751e7455883eb2","source":{"kind":"arxiv","id":"2605.26720","version":1},"attestation_state":"computed","paper":{"title":"Towards Feedback-to-Plan Decisions for Self-Evolving LLM Agents in CUDA Kernel Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiaming Wu, Peng Qu, Yee Hin Chong, YouHui Zhang","submitted_at":"2026-05-26T09:00:09Z","abstract_excerpt":"Large language models (LLMs) have shown strong empirical gains as self-evolving agents for CUDA kernel generation, driven by feedback-conditioned planning across generations. However, how planning decisions attribute and combine heterogeneous feedback signals remains opaque. Standard end-to-end ablations fail to resolve this question, as iterative planning amplifies early perturbations and conflates feedback effects with trajectory-dependent drift.\n  We introduce \\texttt{CUDAnalyst}, a unified analysis layer for controlled, generation-level attribution of planning decisions to feedback compone"},"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":"2605.26720","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-26T09:00:09Z","cross_cats_sorted":[],"title_canon_sha256":"c415c3552e94d7b8e180a847564eee03e872e63854ede5be2e71ef6471d741f1","abstract_canon_sha256":"be3ab33c6a8ac0504adc6beb8e5409e14ac186b6640d9822af5a6d319ed09e40"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:06:09.194794Z","signature_b64":"7ycmQVZaKMTVjc6TqhR3XKji0zh8A4yM1zskfdL5GKRCbrslI2DfStTU65jyJx2i0R4T5ws6fhVD7AW2VNaADA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6f2b3021e6058f6110812c26bce93188b27cc6f4627e41613b751e7455883eb2","last_reissued_at":"2026-05-27T01:06:09.194112Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:06:09.194112Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Feedback-to-Plan Decisions for Self-Evolving LLM Agents in CUDA Kernel Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiaming Wu, Peng Qu, Yee Hin Chong, YouHui Zhang","submitted_at":"2026-05-26T09:00:09Z","abstract_excerpt":"Large language models (LLMs) have shown strong empirical gains as self-evolving agents for CUDA kernel generation, driven by feedback-conditioned planning across generations. However, how planning decisions attribute and combine heterogeneous feedback signals remains opaque. Standard end-to-end ablations fail to resolve this question, as iterative planning amplifies early perturbations and conflates feedback effects with trajectory-dependent drift.\n  We introduce \\texttt{CUDAnalyst}, a unified analysis layer for controlled, generation-level attribution of planning decisions to feedback compone"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.26720","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/2605.26720/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":"2605.26720","created_at":"2026-05-27T01:06:09.194215+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.26720v1","created_at":"2026-05-27T01:06:09.194215+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.26720","created_at":"2026-05-27T01:06:09.194215+00:00"},{"alias_kind":"pith_short_12","alias_value":"N4VTAIPGAWHW","created_at":"2026-05-27T01:06:09.194215+00:00"},{"alias_kind":"pith_short_16","alias_value":"N4VTAIPGAWHWCEEB","created_at":"2026-05-27T01:06:09.194215+00:00"},{"alias_kind":"pith_short_8","alias_value":"N4VTAIPG","created_at":"2026-05-27T01:06:09.194215+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/N4VTAIPGAWHWCEEBFQTLZ2JRRC","json":"https://pith.science/pith/N4VTAIPGAWHWCEEBFQTLZ2JRRC.json","graph_json":"https://pith.science/api/pith-number/N4VTAIPGAWHWCEEBFQTLZ2JRRC/graph.json","events_json":"https://pith.science/api/pith-number/N4VTAIPGAWHWCEEBFQTLZ2JRRC/events.json","paper":"https://pith.science/paper/N4VTAIPG"},"agent_actions":{"view_html":"https://pith.science/pith/N4VTAIPGAWHWCEEBFQTLZ2JRRC","download_json":"https://pith.science/pith/N4VTAIPGAWHWCEEBFQTLZ2JRRC.json","view_paper":"https://pith.science/paper/N4VTAIPG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.26720&json=true","fetch_graph":"https://pith.science/api/pith-number/N4VTAIPGAWHWCEEBFQTLZ2JRRC/graph.json","fetch_events":"https://pith.science/api/pith-number/N4VTAIPGAWHWCEEBFQTLZ2JRRC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N4VTAIPGAWHWCEEBFQTLZ2JRRC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N4VTAIPGAWHWCEEBFQTLZ2JRRC/action/storage_attestation","attest_author":"https://pith.science/pith/N4VTAIPGAWHWCEEBFQTLZ2JRRC/action/author_attestation","sign_citation":"https://pith.science/pith/N4VTAIPGAWHWCEEBFQTLZ2JRRC/action/citation_signature","submit_replication":"https://pith.science/pith/N4VTAIPGAWHWCEEBFQTLZ2JRRC/action/replication_record"}},"created_at":"2026-05-27T01:06:09.194215+00:00","updated_at":"2026-05-27T01:06:09.194215+00:00"}