Context Windows
#content/article
Highlights
The “context window” refers to the entirety of the amount of text a language model can look back on and reference when generating new text plus the new text it generates.
- Represents a “working memory” for the model.
Thinking tokens are treated as output tokens. All inputs and the output from the previous turn is carried forward with the exception of the thinking block, which can be dropped now that Claude has completed the entire use cycle.
Tool uses are treated as output tokens. When tools are being used those tool configurations are included as a part of the original input tokens with the user message
