• Behohippy
    link
    fedilink
    English
    11 year ago

    Any data sets produced before 2022 will be very valuable compared to anything after. Maybe the only way we avoid this is to stick to training LLMs on older data and prompt inject anything newer, rather than training for it.