Pinned Loading

1. N-SUMHSS_LASSO N-SUMHSS_LASSO Public

   Analyzes substance use and mental health treatment facility data from the National Substance Use and Mental Health Services Survey (N-SUMHSS) and its predecessors (N-SSATS, N-MHSS) from 1997-2023.

   Jupyter Notebook 1

2. BVAR_PhD BVAR_PhD Public

   Extends the core methodology from "Large Bayesian Vector Auto Regressions" by Bańbura et al. (2010), focusing on how Bayesian shrinkage enables estimation of high-dimensional VARs without overfitting.

   2 1

3. LLM-constitutional-coding LLM-constitutional-coding Public

   Scalable LLM pipeline for analyzing constitutional text using Q&A-style questioning. Uses topological sorting for dependency scheduling and stateful context injection to handle conditional logic.

   Jupyter Notebook 1

4. Ashley095/Risk-Aware-Calibration-for-Language-Models Ashley095/Risk-Aware-Calibration-for-Language-Models Public

   This project implements temperature scaling calibration for large language models on the HellaSwag multiple-choice question answering task. We evaluate four different models (RoBERTa, DeBERTa, FLAN…

   Python 1