implementation detail · filed under model architecture
Top-scoring compressed-block selection
Each query selects the highest-scoring compressed blocks in Qwen Sparse Attention.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
each query selects the highest-scoring compressed blocks
usedunclearin Qwen3.8-Flash-NextQwen
Filed alongside
Other methods under model architecture :: token mixer :: sparse attention :: block selection.