tasksource/tasksource-jev-typed-decisions
Viewer • Updated • 4.92M • 3.6k • 16
Bo Wang
bowang0911
AI & ML interests
None yet
Recent Activity
new activity about 16 hours ago
tasksource/tasksource-jev-typed-decisions:Add filtered-full train configuration (1,305,478 retained decisions) liked a dataset 3 days ago
tasksource/tasksource-jev-typed-decisions liked a dataset 4 days ago
reasoning-core/procedural-pileOrganizations
Add filtered-full train configuration (1,305,478 retained decisions)
🔥 1
#2 opened about 16 hours ago
by bowang0911
liked a
dataset 3 days ago
liked a
dataset 4 days ago
Remove training source and make inference self-contained
#1 opened 4 days ago
by 
bowang0911
liked a
model 6 days ago
published a
model 6 days ago
Use dedicated PAD token to preserve EOS training labels
#2 opened 11 days ago
by bowang0911
Add a 4/16/4 recurrent 1B architecture and 100-step trial results
#1 opened 11 days ago
by bowang0911
updated a
model 11 days ago
published a
model 11 days ago
Add a trainable 6/24/6 recurrent core with iteration-specific KV caches
#4 opened 12 days ago
by bowang0911
Use 36 independent layers and 1280-wide experts for the main architecture candidate
#3 opened 12 days ago
by bowang0911
updated a
model 12 days ago
Simplify to hybrid GQA and DeepSeek MoE with a 128K context target
#2 opened 12 days ago
by bowang0911
Replace mHC with single-stream pre-norm residual connections
#1 opened 12 days ago
by bowang0911
liked a
Space 28 days ago
liked a
model about 1 month ago
published a
model about 1 month ago
updated a
dataset about 1 month ago
published a
dataset about 1 month ago