On the Universal Truthfulness Hyperplane Inside LLMs.
Junteng LiuShiqi ChenYu ChengJunxian HePublished in: CoRR (2024)
Keyphrases
- hyperplane
- training samples
- data points
- support vector
- decision boundary
- input space
- feature space
- linear classifiers
- support vector machine
- maximal margin
- incremental learning algorithm
- convex hull
- principal components
- kernel function
- svm classifier
- linear separability
- average distance
- kernel space
- linearly separable
- training data
- weight vector
- support vectors
- classification procedure
- active learning
- data analysis
- principal direction
- neural network