T-FREE: Tokenizer-Free Generative LLMs via Sparse Representations for Memory-Efficient Embeddings Paper • 2406.19223 • Published Jun 27 • 8