arXiv API negative search: no paper for NVIDIA DeepGEMM (and a name-collision warning)

TITLE search (ti:) for 'DeepGEMM' returns exactly ONE entry and it is THE WRONG ONE: arXiv 2304.09049 'DeepGEMM: Accelerated Ultra Low-Precision Inference on CPU Architectures using Lookup Tables' (2023), whose authors (Darshan C. Ganji, Saad Ashfaq, Ehsan Saboori, Sudhakar Sah, Saptarshi Mitra, MohammadHossein AskariHemmat, Alexander Hoffman, Ahmed Hassanien, Mathieu Leonardon) are an unrelated CPU lookup-table inference group. That is NOT NVIDIA's DeepSeek-serving FP8/FP4 GEMM library of the same name. Recorded as a first-class gotcha rather than a footnote because a title-only search returning a plausible-looking same-name paper is exactly the mechanism that produces a confident wrong citation - the same shape as the repo's PMLR-volume trap and its duplicated-title trap (DINOv2 at both TMLR and ICLR). Anyone filling the DeepGEMM citation must reject 2304.09049 on author grounds, not on venue grounds.