Skip to content
TheChatArchiveThe Chat Archive
gpt-4.1-2025-04-14· OpenAIHugging Face LMArena Human Preference 140K

Imagine an agentic LLM model that is integrated with an IDE, like VSCode. How do I tell…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Geminigemini-2.5-pro-preview-03-25’s answer.
  1. User

    Imagine an agentic LLM model that is integrated with an IDE, like VSCode. How do I tell it to stop leaving comments in the generated code? For example, I might ask it to generate an algorithm that integrates within the current project I might have. Problem is, instead of re-writing, it leaves comments like:

    function foo(): /* code is same as before */

    function foo2(): /* keep code as before */

    How do i tell it to stop repeating this behavior, in a single concise instruction?

  2. Assistant

    You can give the agent a clear, concise instruction like:

    "When generating or modifying code, do not leave any comments like 'code is same as before' or 'keep code as before'. Only output the actual code, fully written out, with no placeholder comments."

    Or even more succinctly:

    "Output the complete code without any placeholder comments or notes—write out all code explicitly."

    If you want a single sentence to paste as a system prompt or instruction:

    "Always output the full, explicit code with no placeholder comments or notes such as 'code is same as before'—write out all code in full."

    This should make your intent clear to the agentic LLM.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.