Being overly trained on comments in documentation could be one of the reasons why Claude models write frustrating comments (which often manifest as written just

Being overly trained on comments in documentation could be one of the reasons why Claude models write frustrating comments (which often manifest as written justifications of how the code satisfies the prompt). A comment in a tutorial is going to be geared to explaining how the code relates back to the tutorial task or restating what the code does, rather than documenting the "why" of surprising code.