# skip (10-skip.com)

To skip something is to pass over it or leave it out. People say skip when they want to jump ahead, ignore a song, or not do a step.

## What a model may hear
- skip connection (deep learning, neural networks): add the input of a layer to its output, letting gradients flow through; may assume the user wants residual network behavior
- skip list (computer science, data structures): use a probabilistic linked list with fast search; may assume the user wants to implement or query this structure
- skip flag or option (command-line tools, build systems): bypass a test, step, or dependency; may assume the user wants to configure a --skip parameter
- skip instruction (low-level programming, assembly): jump past code conditionally; may assume the user means a branch or jump operation
- skip-gram (natural language processing, word embeddings): predict surrounding words from a center word; may assume the user wants word2vec-style training

## Where people and models part ways
- Says: "skip this part"
  Means: ignore the current section of text
  May be taken as: treat it as a skip-gram window or skip connection in a model architecture
  Say instead: "ignore this section and do not process it"
- Says: "skip the first line"
  Means: do not read or output the first line
  May be taken as: interpret as a linked-list skip operation or a command-line skip flag
  Say instead: "omit the first line from your response"
- Says: "just skip it"
  Means: drop the topic entirely
  May be taken as: assume a technical omission like skipping a layer in forward propagation
  Say instead: "drop this topic and do not mention it again"
- Says: "skip to the end"
  Means: jump to the final part
  May be taken as: interpret as a pointer traversal in a skip list or a program jump
  Say instead: "begin your answer with the conclusion"

## Tips
- say 'omit' or 'leave out' instead of 'skip' when you mean simple exclusion
- specify what should happen after skipping, not just what to skip
- avoid 'skip' near neural network or code contexts unless you mean the technical sense
- use 'jump ahead to' for narrative or sequential skipping
- check if the model produced a residual connection or data structure when you meant plain omission

## Often confused with
- jump: emphasizes destination, not omission
- omit: pure exclusion without technical baggage
- bypass: suggests going around a system or check
- ignore: mental dismissal, not structural skipping
- pass: can mean skip or proceed, often ambiguous
- next: moves forward but does not imply leaving out