specific method · filed under model architecture
Reindex Mode
A CSA2 mode that preserves cache sharing while allowing selected entries to change across layers.
Also called CSA2 Reindex Mode.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- core
How sources treat it
One count per evidence span, weakest treatment to strongest.
optional 1used 1core 1
Documented in
Evidence
3 spans quoted from the sources, strongest treatment first.
Reindex Mode preserves cache sharing while allowing the selected entries to change across layers.
coreunclearin DeepSeek-V4.1-FlashDeepSeek
the first layer operates in Reindex Mode, and the remaining three layers operate in Reuse Mode.
usedunclearin DeepSeek-V4.1-FlashDeepSeek
The layer reuses the most recent available main KV from a preceding layer together with its corresponding indexer K.
optionalinference servingin DeepSeek-V4.1-FlashDeepSeek
Filed alongside
Other methods under model architecture :: token mixer :: sparse attention :: sparse attention indexer.
Compressed Sparse Attention 2Hierarchical Sparse IndexerIndexCacheIndexShareLightning IndexerReuse ModeDense Warm-up StageFine-grained token selectionIndex BranchIndexer WarmupReuse QSA index selection across speculative decoding stepsAverage poolingBlock-causal scoringCompressed lightweight indexerCross-stage shared-state management for attention reuseDetached indexer-input optimizationIndex Branch outputIndex Branch value headIndexPoolMQA indexerSingle-head index key