Jayachandran, D., & Blum, J. J. (2024, April), A Large Language Model Pipeline to Automate the Solution of Competitive Programming Problems  Paper presented at ASEE Mid-Atlantic Section Spring Conference, George Washington University, District of Columbia. 10.18260/1-2--45702

\bibitem{asee_peer_45702}
  Jayachandran, D., \& Blum, J. J. (2024, April), \emph{A Large Language Model Pipeline to Automate the Solution of Competitive Programming Problems}
  Paper presented at ASEE Mid-Atlantic Section Spring Conference, George Washington University, District of Columbia.
  10.18260/1-2--45702

Devang Jayachandran and Jeremy Joseph Blum.  "A Large Language Model Pipeline to Automate the Solution of Competitive Programming Problems".  ASEE Mid-Atlantic Section Spring Conference, George Washington University, District of Columbia, 2024, April.  ASEE Conferences, 2024.  https://peer.asee.org/45702 Internet.  23 Jun, 2024

\bibitem{asee_peer_45702}
  Devang Jayachandran and Jeremy Joseph Blum.
  "A Large Language Model Pipeline to Automate the Solution of Competitive Programming Problems".
  \emph{ASEE Mid-Atlantic Section Spring Conference, George Washington University, District of Columbia, 2024, April}.
  ASEE Conferences, 2024.
  https://peer.asee.org/45702 Internet.  23 Jun, 2024

@INPROCEEDINGS{asee_peer_45702
author = "Devang Jayachandran and Jeremy Joseph Blum"
title = "A Large Language Model Pipeline to Automate the Solution of Competitive Programming Problems"
booktitle = "ASEE Mid-Atlantic Section Spring Conference"
year = "2024"
month = "April"
address = "George Washington University, District of Columbia"
publisher = "ASEE Conferences"
note = {https://peer.asee.org/45702}
number = {10.18260/1-2--45702}
}

TY  - CPAPER
AB  - The recent rapid introduction of large language models has enabled new black box approaches to optimize the application of these models for various scenarios. GPT-4 is a multimodal large language model introduced by OpenAI which can answer complex questions, analyze nuanced data, and solve complicated programming problems. The performance of GPT is dependent upon the provided prompt and hyperparameters. This paper explores the effect of minor variations in system prompt and parameters including temperature and top-p for code generation and code accuracy for competitive programming tasks. Temperature controls the amount of randomness in the response, with a temperature of zero producing deterministic output.  Top-p controls the cumulative probability distribution for tokens considered for the next output token.  Based on the results, we propose approaches to optimize system prompts for code-generation and parameter values to improve the correctness of code. In addition, we propose a pipeline that utilizes these enhancements to effectively solve algorithmic puzzles common in computer science education, in addition to complex contest programming problems.
AU  - Devang Jayachandran
AU  - Jeremy Joseph Blum
CY  - George Washington University, District of Columbia
DA  - 2024/04/19
PB  - ASEE Conferences
TI  - A Large Language Model Pipeline to Automate the Solution of Competitive Programming Problems
UR  - https://peer.asee.org/45702
DO  - 10.18260/1-2--45702
ER  -