
- get_pride_preditions.py to reproduce the results for PriDe baseline
- qk_stat_run.py and att_stat_run.py to get the results on different option-representative tokens used for getting scores
- heads_ablation.py and heads_ablation_random.py are used to reproduce results of ablation expermiment
- get_logit_lens.py is used to reproduce the logit lens expermiments results
- demonstrations.py contains dictionary with examples we used for n-shot prompting strategies
- Example.ipynb containts Jupyter Notebook with code to reproduce our main experiments 