/whypy

an llm-botherer collection

Primary LanguageHTML

whypy

an iterative LLM-botherer

iteratively invoking Large Language Models (LLM) using Ollama via langchain_community package

Scripts

1. _loop.py

iteratively invoke an LLM with different prompts

2. explainyourself.py

iteratively prompts an LLM to explain itself. saves the responses in csv and txt file formats.

Requirements

  • Python 3.x
  • Packages: argparse, pandas, langchain_community, langchain, langchain_core

install the required packages using pip:

pip install argparse pandas langchain_community

Usage

  1. clone the repo or copy the script to your local machine.
  2. run the script with the desired number of iterations:
python loop.py [iterations]

(if no iterations are given, the loop scripts default to 10)

output files will be saved in the outputs/pitching directory with a filename format that includes the model name, temperature, and a timestamp.

python explainyourself.py

output files are saved in the outputs/* directories with filenames that include the model name and a timestamp.