OpenReview api2 /notes/search responses for the 16 duplicate-slug papers (venue + venueid per forum)

The VENUE evidence for the duplicate-slug consolidation, gathered on 2026-10-05 with the api2 notes/search endpoint, which nests every content field under a 'value' key (api1 does not) and returns DBLP-mirror records alongside genuine conference decision records. THE MECHANICAL POINT THAT MATTERS MOST, and it is a trap of the same family as the PMLR-volume one this repo already logs: a 'Rejected_Submission' venueid and an acceptance venueid appear in the SAME SEARCH RESULT SET for the SAME TITLE with the SAME AUTHORS, and a naive parse takes whichever it reads first. For Mamba the search returns three corroborated records - 'CoRR 2023' / dblp.org/journals/CORR/2023 (forum u7TcGRvHEr, the arXiv deposit), 'COLM' / colmweb.org/COLM/2024/Conference (forum tEYskw1VY2, the ACCEPTANCE), and 'Submitted to ICLR 2024' / ICLR.cc/2024/Conference/Rejected_Submission (forum AL1fq05o7H, the REJECTION). RetNet has the same shape with only the rejection and no acceptance. PER-PAPER RESULT, each title matched on title AND author overlap, never on title alone: Mamba -> COLM 2024 accepted (and the ICLR 2024 rejection is not a venue); S4 -> 'ICLR 2022 Oral' / ICLR.cc/2022/Conference; Splitwise -> 'ISCA 2024' / dblp.org/conf/ISCA/2024 plus the arXiv deposit; SGLang -> 'NeurIPS 2024 poster' / NeurIPS.cc/2024/Conference and DBLP 'NeurIPS 2024'; DISTFLASHATTN -> 'COLM' / colmweb.org/COLM/2024/Conference; Hyena Hierarchy -> 'ICML 2023 OralPoster' / ICML.cc/2023/Conference; Mamba-2 -> 'ICML 2024' / dblp.org/conf/ICML/2024 plus arXiv comment 'ICML 2024'; RetNet -> ICLR 2024 Rejected_Submission ONLY, so no peer-reviewed venue; RWKV -> 'EMNLP 2023 Findings' (S2 DOI 10.18653/v1/2023.findings-emnlp.936); Titans -> 'NeurIPS 2025 poster' / NeurIPS.cc/2025/Conference; BASED -> 'ICML 2024 Spotlight' / ICML.cc/2024/Conference, PLUS a separate 'ES-FoMo-II 2024 Oral' workshop record (ICML.cc/2024/Workshop/ES-FoMo-II) that is a different submission of the same title and must not be mistaken for the main-track entry; Kimi Linear and Lightning Attention-2 -> no conference record of any kind for these titles. TWO TITLE SEARCHES ALSO PRODUCED BAIT WORTH NAMING. A search for 'AIBrix' returns only notes with null titles and null venues ('Author Response to Reviewer') - the name is too short for the search index to match on, so the apparent result set is empty of evidence rather than evidence of absence. A search for 'TurboRAG' returns 'ICLR 2025 Conference Withdrawn Submission' AND 'ACL ARR 2025 May Submission' with the SAME title, and the ARR record is the one that turned out to be a real publication - see w6p-aclanthology-2025-emnlp-main-334-turborag. The direct /notes?forum= endpoint is challenge-walled from this host (HTTP 403 ChallengeRequiredError), so forum threads and decision notes could not be read; the search endpoint is the only one that answers.