SwiGLU-style gated feed-forward projection

Read the equation

Two learned maps project x into the intermediate width. The configured activation, SiLU in these dossiers, is applied to the gate projection; its result multiplies the up-projection componentwise. A down-projection maps that gated intermediate vector back to the residual width. This formula does not itself establish the numerical exactness of a substituted linear layer.

Terms & definitions

How it works

Read the definition from its right-hand expression to the quantity on the left. Its surrounding source records the intended input, state, or parameter context. Where the source leaves an operation unspecified, this page preserves that gap.

Assumptions & limits

Source recovery establishes that the expression appears in this record. Its mathematical validity, implementation behavior, and interpretation must be assessed separately. See the linked source and research notes for the checks available for this version.

Original context

self.act_fn = ACT2FN[config.hidden_act]

    def forward(self, x):
        down_proj = self.down_proj(
            self.act_fn(self.gate_proj(x)) * self.up_proj(x))
        return down_proj

    def to_exact(self, act_bits=16):

down_proj = self.down_proj(
            self.act_fn(self.gate_proj(x)) * self.up_proj(x))

Related builds