21–30 of 124 posts
I think this is a really cool idea and would be something I would use. Is there some way to translate short phrases like "drehte sich um" all together rather than word by word?
I know I’ve seen a lot of complaints on the Latin sub Reddit about GPT generating really bad English sounding Latin.
However, for learners coming from English, this may actually be a benefit as you can get exposed to reading way faster. And ultimately reduce the time until you can read Kafka and Goethe…
Bravo!
Earlier quoted context omitted.
That would be really helpful because taken out of context, single words translated to English will often produce very confusing results. My suggestion would be to ditch the google links and provide a properly translated version that is revealed when hovering over a word/phrase. Such a translation could be crowd sourced.
Yea, I can definitely improve that! That sounds like better way to do it. Thanks!
GPT-3.5 is already very good at sense disambiguation
However, it looks like a lot of this is auto generated/translated, making it sound unnatural or sometimes just wrong.
For example:
I looked at this example story [1] and the practice question is slightly bad grammar. It should be "Wer ging im Wald spazieren?", not "Wer ging spazieren im Wald?".
Or: "Sie lachten und hatten eine gute Zeit [...]" --> This seems like a bad translation of "had a good time", which doesn't translate directly to German like this.
Would definitely not recommend this overall.
[1] https://webbu.app/l/german/story/ein-unerwarteter-spaziergan...
Intermediate German student here. I haven’t studied German in about seven years. I found the example story surprisingly easy to understand. I wonder if that’s because it was generated by an LLM, which is mostly trained on English and is generating, basically thinly veiled English story in German. I know I’ve seen a lot of complaints on the Latin sub Reddit about GPT generating really bad English sounding Latin. Howev…
If this is using LLMS then it would be great to expand to as many languages as possible because anthing outside of large languages are very poorly served by existing language learning apps. Danish (from english) would be my preference, but I imagine you could have a user configurable pair (I understand the situation is much, much more dire if you don't even presently understand a large language english,spanish,chines…
This is awesome! I've been casually learning German for 200+ days now, so this will make a great addition to my practice. Are the short stories generated by LLMs too?
Would it be possible to add different translation engines, like DeepL or ChatGPT? (ChatGPT in particular is often much better than all the alternatives.)