implementation detail · filed under model architecture
Compressed lightweight indexer
A lightweight indexer that selects important context at micro-block granularity.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- core
How sources treat it
One count per evidence span, weakest treatment to strongest.
core 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
uses a compressed lightweight indexer to select the important context at micro-block granularity
coremodel architecturein Qwen3.8-Flash-NextQwen
Filed alongside
Other methods under model architecture :: token mixer :: sparse attention :: sparse attention indexer.
Compressed Sparse Attention 2Hierarchical Sparse IndexerIndexCacheIndexShareLightning IndexerReindex ModeReuse ModeDense Warm-up StageFine-grained token selectionIndex BranchIndexer WarmupReuse QSA index selection across speculative decoding stepsAverage poolingBlock-causal scoringCross-stage shared-state management for attention reuseDetached indexer-input optimizationIndex Branch outputIndex Branch value headIndexPoolMQA indexerSingle-head index key