Code

Syntax-highlighted code blocks via Prism, with light and dark theme support.

---

Inline Code

Use single backticks: const x = 42.

---

Code Blocks

Specify the language after the opening triple backticks. Wrap in <Code> for a numbered, captioned, cross-referenceable block:

import torch.nn as nn

class TransformerBlock(nn.Module):
    def __init__(self, d_model, n_heads, d_ff):
        super().__init__()
        self.attn = nn.MultiheadAttention(d_model, n_heads)
        self.ff = nn.Sequential(nn.Linear(d_model, d_ff), nn.GELU(), nn.Linear(d_ff, d_model))
        self.norm1 = nn.LayerNorm(d_model)
        self.norm2 = nn.LayerNorm(d_model)

    def forward(self, x):
        x = self.norm1(x + self.attn(x, x, x)[0])
        x = self.norm2(x + self.ff(x))
        return x

def train_step(model, batch, optimizer):
    optimizer.zero_grad()
    loss = model(batch)
    loss.backward()
    optimizer.step()
    return loss.item()

Reference them inline: see and .

---

Supported Languages

Language	Tag	Example	Description
Python	python	def foo(): pass	ML, data science, scripting
JavaScript	js	const x = 42;	Frontend, Node.js, React
TypeScript	ts	let x: number = 42;	Typed JavaScript
Bash	bash	echo "hello"	Shell scripts, CLI commands
C / C++	c / cpp	int main() {}	Systems, CUDA kernels
Java	java	System.out.println()	Enterprise, Android
Rust	rust	fn main() {}	Systems, performance-critical
YAML / JSON	yaml / json	key: value	Configuration files
LaTeX	latex	\frac{a}{b}	Math typesetting
SQL	sql	SELECT * FROM t	Database queries

Full list at Prism supported languages.

---

Algorithms

The <Algorithm> component renders a bordered pseudocode block with a numbered caption:

Input: learning rate $\eta$, initial parameters $\theta_0$, dataset $\mathcal{D}$

1. for $t = 1, 2, \ldots, T$ do
2.     Sample mini-batch $\mathcal{B} \subset \mathcal{D}$
3.     $g_t \leftarrow \frac{1}{|\mathcal{B}|} \sum_{(x,y) \in \mathcal{B}} \nabla_\theta \ell(f_\theta(x), y)$
4.     $\theta_t \leftarrow \theta_{t-1} - \eta \cdot g_t$
5. end for

Output: trained parameters $\theta_T$

Reference: see .