Skip to main content
Search

Mir Tafseer Nayeem

2 related articles

Subscribe to updates
  1. arXiv

    Across nearly half a million names and 12 tokenizers, whether a name becomes a single token shifts concept accessibility in fellowship, hiring, clinical, and lending judgments

    The study introduces NameTrace, measures direct lexical support for names across nearly half a million first names and 12 LLM-associated tokenizers, and, on atomic versus short-fragmented names matched within the same race/ethnicity–gender strata, finds systematically higher task-aligned concept accessibility for atomic names across fellowship, hiring, clinical assessment, and lending; the differences persist in all eight strata, transfer to unseen names, and hidden-state interventions along the measured task directions shift later constrained choices.

  2. arXiv

    Auditing 1,813 AmE–BrE variant pairs across the pipeline: six pretraining corpora, 21 post-training datasets and ten checkpoints all lean American English

    The study builds a curated resource of 1,813 matched American English–British English variant pairs and introduces DiAlign, a training-free method for estimating regional alignment, then uses them to audit six pretraining corpora, 21 post-training datasets, nine tokenizers, ten model checkpoints and generations under two prompting conditions, finding that American English is systematically favored across data exposure, representation and generation, and that British-English prompting only partially shifts this default.

You’re all caught up