Skip to content

reproduce some of the content #22

@Kenwwww

Description

@Kenwwww

I'd like to reproduce some of the content. Regarding some details not mentioned in the report, could you please offer some guidance?

A large amount of synthetic inference data was mixed in the pre-training. What is the format of this part of the data? Is it necessary to add special Tokens?

Is it okay like this?

{"text":"query... ? reasoninganswer..." }

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions