That's interesting. I am curious about the extent of the training data in these models.
I asked Kimi K2 for an account of growing up in my home town in Scotland, and it was ridiculously accurate. I then asked it to do the same for a similarly sized town in Kerala. ChatGPT suggested that while it was a good approximation, K2 got some of the specifics wrong.
I asked Kimi K2 for an account of growing up in my home town in Scotland, and it was ridiculously accurate. I then asked it to do the same for a similarly sized town in Kerala. ChatGPT suggested that while it was a good approximation, K2 got some of the specifics wrong.