Pinned Loading

1. llm-steering-audit llm-steering-audit Public

   Code for "White-Box Sensitivity Auditing with Steering Vectors"

   Python 3

2. llm-censorship-steering llm-censorship-steering Public

   Code for 'Steering the CensorShip: Uncovering Representation Vectors for LLM "Thought" Control'

   Python 12 1

3. gender-bias-steering gender-bias-steering Public

   Code for the paper "Unsupervised Concept Vector Extraction for Bias Control in LLMs"

   Jupyter Notebook 2 1

4. balanced-adversarial-training balanced-adversarial-training Public

   Python 2

5. composed-debiasing composed-debiasing Public

   Jupyter Notebook 1

6. automatic-paraphrase-dataset-augmentation automatic-paraphrase-dataset-augmentation Public

   Code and data for automatic paraphrase dataset augmentation.

   Jupyter Notebook 11