The Incumbent Did the Work: What the Simple-Baseline Results in LLM-Guided Program Search Settle, Why the Dispute Over Archives, Diversity and Test-Time Training Is Fought on Benchmarks That Cannot Rank It, and the Comparisons That Would

APA

Mahendrakar, P. (2026). The Incumbent Did the Work: What the Simple-Baseline Results in LLM-Guided Program Search Settle, Why the Dispute Over Archives, Diversity and Test-Time Training Is Fought on Benchmarks That Cannot Rank It, and the Comparisons That Would [Preprint]. Life of Research. https://doi.org/10.5281/zenodo.23684576

Copy APA

BibTeX

@misc{lor1482026,
  title        = {The Incumbent Did the Work: What the Simple-Baseline Results in LLM-Guided Program Search Settle, Why the Dispute Over Archives, Diversity and Test-Time Training Is Fought on Benchmarks That Cannot Rank It, and the Comparisons That Would},
  author       = {Pranay Mahendrakar},
  year         = {2026},
  publisher    = {Life of Research},
  doi          = {10.5281/zenodo.23684576},
  url          = {https://lifeofresearch.com/preprints/148},
  keywords     = {LLM-guided program search, evolutionary search, test-time compute, AlphaEvolve, hill climbing, test-time training, evaluation methodology, mathematical discovery},
  note         = {LOR-00148}
}

Copy BibTeX More on the record page