Retaining Key Information under High Compression Ratios: Query-Guided Compressor for LLMs

Open in new window