Arxiv Assistant

I'm currently rewriting this using Streamlit to allow me to give users more customizability. Here are a few things that I'm trying to improve over the last iteration:

• Pick which LLM you want to infer from
• From the UI, select how many papers you want to retrieve
• Disable or enable retrieving more papers

Run the app

Run streamlit run app.py. Configuration (API keys, models, etc.) is within the UI.

Known issues

• If your "Maximum input tokens" is smaller than the size of one abstract, there can be issues.

Screenshots