Characterizing Prompt Compression Methods for Long Context Inference

Open in new window