Llama 3 8B · Attention update

32 independent transformer blocks › Attention

Attention update

tensorself.wo(output)

The model-width contribution added to the containing residual stream.

shape: $\Delta_{\mathrm{attn}}$ [B batch, 1 position, 4096 d]

Codemodel.py 121–248 · sku_list.py 19, 188–207

        self.attention_norm = RMSNorm(args.dim, eps=args.norm_eps)
        self.ffn_norm = RMSNorm(args.dim, eps=args.norm_eps)

    def forward(
        self,
        x: torch.Tensor,
        start_pos: int,
        freqs_cis: torch.Tensor,
        mask: Optional[torch.Tensor],
    ):
        h = x + self.attention(self.attention_norm(x), start_pos, freqs_cis, mask)
        out = h + self.feed_forward(self.ffn_norm(h))
        return out

model.py 121–127

        self.wo = RowParallelLinear(
            args.n_heads * self.head_dim,
            args.dim,
            bias=False,
            input_is_parallel=True,
            init_method=lambda x: x,
        )

model.py 189–190

        output = output.transpose(1, 2).contiguous().view(bsz, seqlen, -1)
        return self.wo(output)

sku_list.py 19

LLAMA3_VOCAB_SIZE = 128256

sku_list.py 188–207

def llama3_base_models() -> list[Model]:
    return [
        Model(
            core_model_id=CoreModelId.llama3_8b,
            description="Llama 3 8b model",
            huggingface_repo="meta-llama/Llama-3-8B",
            arch_args={
                "dim": 4096,
                "n_layers": 32,
                "n_heads": 32,
                "n_kv_heads": 8,
                "vocab_size": LLAMA3_VOCAB_SIZE,
                "ffn_dim_multiplier": 1.3,
                "multiple_of": 1024,
                "norm_eps": 1e-05,
                "rope_theta": 500000.0,
                "use_scaled_rope": False,
            },
            pth_file_count=1,
        ),

Sources3

Contextual evidence

The block adds the attention update to its input before normalization.

block-residuals · meta-llama/llama3 / llama/model.py
Pinned revision a0940f9cf706 ↗ · lines 236–248

Read code excerpt

        self.attention_norm = RMSNorm(args.dim, eps=args.norm_eps)
        self.ffn_norm = RMSNorm(args.dim, eps=args.norm_eps)

    def forward(
        self,
        x: torch.Tensor,
        start_pos: int,
        freqs_cis: torch.Tensor,
        mask: Optional[torch.Tensor],
    ):
        h = x + self.attention(self.attention_norm(x), start_pos, freqs_cis, mask)
        out = h + self.feed_forward(self.ffn_norm(h))
        return out

Contextual evidence

A bias-free learned output projection transforms the merged field at model width.

attn-output · meta-llama/llama3 / llama/model.py
Pinned revision a0940f9cf706 ↗ · lines 121–190

Read code excerpt

        self.wo = RowParallelLinear(
            args.n_heads * self.head_dim,
            args.dim,
            bias=False,
            input_is_parallel=True,
            init_method=lambda x: x,
        )

… lines 128–188 omitted …

        output = output.transpose(1, 2).contiguous().view(bsz, seqlen, -1)
        return self.wo(output)

Contextual evidence

Meta’s original Llama 3 8B base catalogue entry supplies dimensions, head counts, vocabulary, MLP rounding settings, epsilon and unscaled RoPE settings.

config · meta-llama/llama-models / models/sku_list.py
Pinned revision 0e0b8c519242 ↗ · lines 19–207

Read code excerpt

LLAMA3_VOCAB_SIZE = 128256

… lines 20–187 omitted …

def llama3_base_models() -> list[Model]:
    return [
        Model(
            core_model_id=CoreModelId.llama3_8b,
            description="Llama 3 8b model",
            huggingface_repo="meta-llama/Llama-3-8B",
            arch_args={
                "dim": 4096,
                "n_layers": 32,
                "n_heads": 32,
                "n_kv_heads": 8,
                "vocab_size": LLAMA3_VOCAB_SIZE,
                "ffn_dim_multiplier": 1.3,
                "multiple_of": 1024,
                "norm_eps": 1e-05,
                "rope_theta": 500000.0,
                "use_scaled_rope": False,
            },
            pth_file_count=1,
        ),