hotchpotch/query-context-pruner-multilingual-Qwen3-4B
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Jul 6, 2025License:mitArchitecture:Transformer0.0K Open Weights Warm

The hotchpotch/query-context-pruner-multilingual-Qwen3-4B is a 4 billion parameter, multilingual Qwen-3 series model developed by Yuichi Tateno. It is specifically designed to identify and remove query-irrelevant context chunks in Retrieval-Augmented Generation (RAG) systems. This model excels at generating high-quality teacher labels for training more efficient, lightweight context pruning models, supporting 20 languages with varying performance levels.

Loading preview...