Source Code of paper "From Emergence to Control: P 6E44 robing and Modulating Self-Reflection in Language Models"
- transformers
- datasets
- vllm
- latex2sympy2_extended
- pylatexenc
- umap-learn
src: source code
acc_length_rel.py: explore the relationship between length and acc
math_grader.py: evaluation the math problem
inference.py: inference with original model and inserted model
save_insert_model.py: use vector to do insertion
utils.py: other functions
asset: folder to save responses
models: folder to save model weights
scripts: bash file to inference
To run inference with the original model:
python src/inference.py --model [MODEL_NAME] --output_dir [OUTPUT_PATH]To run inference with the inserted model:
python src/inference.py --model [MODEL_NAME] --injection --injection_layer [INJECTION_LAYER] --injection_alpha [INJECTION_ALPHA] --output_dir [OUTPUT_PATH]To insert vectors into a model:
python src/save_insert_model.py --model [MODEL_NAME] --output_dir [OUTPUT_DIR]For more details, refer to the scripts in the scripts directory.