You are an issue-hint writer. You need to formulate some hints for detecting certain issues, based on the description in <issue> below. 

<rules>
- The hints should have three different levels, from very general to very specific. 
- For “general hint”, you should only tell the general issue, without any specific information.
- For “medium hint”, you could tell which files are involved in this issue  + general hint. Make sure it contains the information from the general hint.
- For “specific hint”, you should tell which files are involved in this issue clearly,  as well as a part of context information.  Make sure it contains the information from the general and medium hints.
- You should never tell the agent the full context of the issue, only give it hints, so we can test to what degree the agent detects the issue and tell us evidence by themselves.
- Your word used in hints should be precise, pay attention to the word choice, and try to avoid misleading or ambiguous words. You should make sure you have a clear understanding of the issue before writing the hints.
- The hints should be clear and concise, try to cover all the issues mentioned and avoid unnecessary information.
</rules>

<example>
{
    <issue>
    {
        title: What is the provenance of these benchmark datasets?}
        content: 
            {
                Firstly, check the English proverbs dataset: https://github.com/google/BIG-bench/tree/main/bigbench/benchmark_tasks/english_proverbs
                ```
                    task can help to draw conclusions about human-like understanding and
                    ## Data source
                    ## References
                ```
                There is nothing in the data source section.
            }
        involved: {
            name: README.md, 
            context: {
                ```
                    task can help to draw conclusions about human-like understanding and
                    Data source
                    References
                ```
            }
        }
    }
    </issue>

    output:
    {
        general: “section with empty content”
        medium:  “a section with empty content in markdown file”
        specific: “a sub-heading section in README.md with empty content”
    }

</example>

------

Now let’s begin

<issue>
<ISSUE>
</issue>

write your hints here in JSON format: