32 independent transformer blocks › SwiGLU
Gate features
×tensorself.w1(x)
Learned feature combinations before the gate nonlinearity.
- shape
- [B batch, T positions, 14336 d_ff]
Codemodel.py 208–219
self.w1 = ColumnParallelLinear(
dim, hidden_dim, bias=False, gather_output=False, init_method=lambda x: x
)
self.w2 = RowParallelLinear(
hidden_dim, dim, bias=False, input_is_parallel=True, init_method=lambda x: x
)
self.w3 = ColumnParallelLinear(
dim, hidden_dim, bias=False, gather_output=False, init_method=lambda x: x
)
def forward(self, x):
return self.w2(F.silu(self.w1(x)) * self.w3(x))Sources1
Contextual evidence
The MLP owns three bias-free projections. Its forward expression applies SiLU to w1(x), multiplies w3(x) element-wise and projects through w2.
mlp · meta-llama/llama3 / llama/model.py
Pinned revision a0940f9cf706 ↗ · lines 208–219
Read code excerpt
self.w1 = ColumnParallelLinear(
dim, hidden_dim, bias=False, gather_output=False, init_method=lambda x: x
)
self.w2 = RowParallelLinear(
hidden_dim, dim, bias=False, input_is_parallel=True, init_method=lambda x: x
)
self.w3 = ColumnParallelLinear(
dim, hidden_dim, bias=False, gather_output=False, init_method=lambda x: x
)
def forward(self, x):
return self.w2(F.silu(self.w1(x)) * self.w3(x))