(post deleted by author)
Even generating text in Japanese for reading practice for instance seems so odd to me, you have access to a virtually infinite amount of Japanese for basically any level online, why risk having AI generate something broken?
That happens with any language, even English. LLM are just statistical models, they regurgitate whatever was fed to them. They over use words (“Why Does ChatGPT ‘Delve’ So Much?”: FSU researchers begin to uncover why ChatGPT overuses certain words - Florida State University News), they don’t know if the output is correct or not (it has no concept of “knowing”, even), and they’re programmed to say whatever we want them to say, which is not to say that they’re programmed to say the right things.
In some models, even if it gives you the right explanation and you wrongly correct them, or pose a question in a certain way, it will “say” that they were wrong and wrongly correct themselves, just because that’s what you were looking for.
1 Like
Yeah as has been pointed out in this thread already, a huge issue with these language models in my experience is their inability to answer “I don’t know” or even ask follow-up questions when they lack information. Instead they just write out a plausible-sounding but sometimes completely wrong answer.
And the more niche the topic the worse it gets. That’s why it looks really impressive when you just ask simple questions as the model will usually have ample enough data to reply something coherent. But as soon as you start getting into trickier topics the results can get really poor really quickly.
I usually just ask it what the nuances are between two words. Like “How is 自分 different from 私 or 僕?”
The answer it gave sounded pretty legit to me.
「TL;DR:
Use 私 when being polite and neutral
Use 僕 if you’re male and in a friendly or soft setting
Use 自分 when you’re emphasizing responsibility, identity, or doing something yourself」
There was obviously more to this, but it all made understanding 自分 easier. It even added in about 俺 when explaining 僕.
At the end of the day, I follow-up with many resources. I like Bunpro, but I also find it to be aggravating at times. Without having an actual 日本人 spending time teaching me Japanese, from an area I might go to, every resource has its pros and cons.
If that’s what it said that’s actually a pretty terrible explanation for 自分 IMO. The key for 自分 is that it’s not a first person pronoun, it’s a reflective pronoun, so it can be applied to third persons unlike 私、僕、俺、etc… The closest equivalent in English would be “oneself”. There’s more to it but if I was to give a quick explanation that’s certainly where I would start.
I’ll repeat what I said above and assert that for these very simple questions you’re almost always better off just searching for articles online before rolling dice with AI, for instance:
Amusingly while looking for other articles I found this one which seems to be mostly or entirely AI-generated and of extremely low quality as a result: