We are interested in reconstructing fine-grained scientific feuds using large language models (LLMs). We use group selection theory as our case study.
The repository contains the following elements:
- 📂 data raw data
- 📂 src code used for the analysis and to generate figures
- 📂 figs figures from the paper
- 📂 output stuffs generated by code
- 📂 etc miscellaneous
We do contextual citation analysis (CCA) and stance detection fine-tuning openAI's currie model on annotated datasets from DominikBeese/DidAIGetMoreNegativeRecently and allenai/multicite.