src/PositionalEmbeddings.jl: Sanitize docstrings.

src/PositionalEmbeddings.jl

@@ -171,19 +171,12 @@ where the upper triangle including diagonal is masked (True) and the lower trian
 - 3D boolean array of shape (seq_len, seq_len, 1) where True indicates positions to mask
 
 # Examples
-```jldoctest
+```
 julia> mask = create_causal_mask(3)[:,:,1]
 3×3 Matrix{Bool}:
  1  1  1  # First position can't attend anything
  0  1  1  # Second position can attend to first only
  0  0  1  # Third position can attend to first and second
-
-# Used with attention_mask:
-julia> attention_mask(mask)[:,:,1]
-3×3 Matrix{Float64}:
-   -Inf     -Inf    -Inf
-   0.0      -Inf    -Inf
-   0.0      0.0     -Inf
 ```
 """
 function causal_mask(seq_len::Int)
@@ -204,24 +197,18 @@ Create padding masks for batched sequences of varying lengths. This ensures that
 - 3D boolean array of shape (batch_size, max_len, 1) where True indicates padded positions
 
 # Examples
-```jldoctest
+```
 # For 2 sequences of lengths 2 and 3, padded to length 4:
 julia> mask = create_padding_mask([2, 3], 4)[:,:,1]
 2×4 Matrix{Bool}:
  0  0  1  1  # First sequence: length 2, positions 3-4 are padding
  0  0  0  1  # Second sequence: length 3, position 4 is padding
-
-# Combined with attention_mask:
-julia> attention_mask(mask)[:,:,1]
-2×4 Matrix{Float64}:
-  0.0    0.0   -Inf   -Inf
-  0.0    0.0    0.0   -Inf
 ```
 
 # Usage with Causal Mask
 Padding and causal masks are often combined for batched autoregressive tasks:
 
-```julia
+```
 seq_len = 5
 batch_lengths = [3, 4]
 
@@ -231,7 +218,6 @@ padding = create_padding_mask(batch_lengths, seq_len) # Shape: (2, 5, 1)
 
 # Combine masks which will either prevent attending to future tokens or padding tokens
 combined = causal .| padding
-final_mask = attention_mask(combined)
 
 # final_mask will prevent:
 # 1. Attending to future tokens (from causal mask)