simple-stories-3L16H128D-attention-only-toy-transformer

3-Layer 16-Head Attention-Only Transformer

This is a simplified transformer model with 3 attention layer(s) and 16 attention head(s), hidden size 128, designed for studying attention mechanisms in isolation.

Architecture Differences from Vanilla Transformer

Removed Components:

• No MLP/Feed-Forward layers - Only attention layers
• No Layer Normalization - No LayerNorm before/after attention
• No positional encoding - No position embeddings of any kind

Kept Components:

• Token embeddings
• Multi-head self-attention with causal masking
• Residual connections around attention layers
• Language modeling head (linear projection to vocabulary)

This minimal architecture isolates the attention mechanism, making it useful for mechanistic interpretability research as described in A Mathematical Framework for Transformer Circuits.

Usage

    config_class = LlamaConfig

    def __init__(self, config: LlamaConfig):
        super().__init__(config)
        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size)
        self.layers = nn.ModuleList([AttentionLayer(config) for _ in range(config.num_hidden_layers)])
        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)


model = AttentionOnlyTransformer.from_pretrained('Butanium/simple-stories-3L16H128D-attention-only-toy-transformer')

Training Data

The model is trained on the SimpleStories dataset for next-token prediction.