Skip to content
Adith Santosh

Research

What models mean, and how we’d know

My work sits between natural language understanding and measurement: bias and fairness, explainability, calibration, and the geometry of embeddings.

Natural language understandingLarge language modelsExplainabilityEvaluation

Publications

Published as Adith Santosh Thaniserikaran

  1. [1]

    Phonetic Cues Improve LLM-Based Pun Detection in Short Text

    ACL 2026 Computational Humor Proceedings (CHum 2026)

  2. [2]

    Review-Based Group Recommender Systems Using Deep Learning

    Proceedings of the ACM Colloquium on Advances in Data Science and Computing (CADSCOM)

    Top Paper Award

Positions

  1. Purdue University

    Aug 2025 – Present

    Research Assistant

    Research on NLU, bias and fairness analysis, and explainability for LLMs, with an emphasis on statistical evaluation, calibration and embedding geometry (similarity correlations, cluster separability).

  2. New York University

    Sep 2024 – May 2025

    Research Assistant

    ML surrogate models for OpenFAST wind-turbine simulations, improving dynamic load prediction accuracy by 12% and reducing simulation time with regression and time-series models.

Collaborate on research →Applied projects