Simple Text Detoxification by Identifying a Linear Toxic Subspace in Language Model Embeddings

Re-assign community