Still not open weights?
EuroLLM: LLM made in Europe built to support all 24 official EU languages
481–490 of 629 posts
Re: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#482If multi‑lang is the goal, why not translate the output of the big labs?
Languages don't operate one to one.
Re: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#483Earlier quoted context omitted.
AFAIK, they are trying to get Frisian added to the "European Charter for Regional or Minority Languages", not the official language list. They get certain recognition, but they are not official in Europe. For example, just from Spain there are 13 languages on that list.
To be fair to the Frisians, there are around 40,000-80,000 native Irish speakers and 500,000+ native Frisian speakers.
Unfair comparison, imo. Irish (Gaelic) is a language which was intentionally suppressed for centuries.
Re: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#484Earlier quoted context omitted.
The issue starts, when an LLM's transfers knowledge between languages, even though that knowledge is not correct in that language. I have seen this with e.g. ChatGPT answers regarding laws for example where it refers to US laws when asked in German, which are obviously not relevant.
> The issue starts, when an LLM's transfers knowledge between languages, even though that knowledge is not correct in that language. I have seen this with e.g. ChatGPT answers regarding laws for example where it refers to US laws when asked in German, which are obviously not relevant. There is no necessary correlation between language and the correct set of laws to reference. The language of the question (or the answ…
When you ask an LLM (in German) without further clarifying your location I expect it to refer to German (or Austrian/Swiss) laws.
For most questions it does this pretty well (e.g. asking for the legal age to drink). However once the answer becomes more complex it starts to halucinate very quickly. The fact that some of the hallucinations are just translated US laws makes me think that the knowledge transfer between languages is probably not helping in instances like this.
Re: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#485Earlier quoted context omitted.
This completely ignores the history of published writing in Ulster Scots going back centuries.
Part of the issue some people take with Ulster-Scots is that the current official 21st century literature doesn't read anything like the historic literature, which English speakers can easily read and understand. It's often made up of slang terms and archaic spelling, in an attempt to be as different as possible to English. Native speakers have complained that official documents and signage in Ulster-Scots are incomp…
> the current official 21st century literature doesn't read anything like the historic literature
Does modern English read like historical English?
> Native speakers have complained that official documents and signage in Ulster-Scots are incomprehensible to them.
Sure, there are tonnes of issues with the "officialisation" of any language but the fact that there are "native speakers" involved in the debate strongly suggests it wasn't all just made up for political reasons, which was the point I was responding to.
Re: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#486Earlier quoted context omitted.
Is there any dialect of Arabic which you can understand without too much effort? How much do you consider Maltese its own language (as opposed to a dialect of Arabic)?
From what I have heard, Lebanese Arabic is the closest, and still pretty far. Passable conversation is possible. Maltese is definitely its own language. Arabic roots are there (theres a Semitic joke in there ) but it isn't arabic anymore. Its written left to right with a variant of the english alphabet.
Writing RTL or LTR and alphabet alone don't make a language different.
Hindi and Urdu are 90% the exact same language, and are mutually inteligible (Urdu speaker and Hindi speaker can have complete full conversation with each other) but each is written differently (one LTR the other RTL) and with different alphabets
Re: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#487Earlier quoted context omitted.
AFAIK, they are trying to get Frisian added to the "European Charter for Regional or Minority Languages", not the official language list. They get certain recognition, but they are not official in Europe. For example, just from Spain there are 13 languages on that list.
To be fair to the Frisians, there are around 40,000-80,000 native Irish speakers and 500,000+ native Frisian speakers.
Irish is an official language of Ireland (there is signage and instructions in Irish up and down the Republic) , Frisian is not an official language of the Netherlands to the best of my knowledge
Irish is certainly not a robust vigorous language but your 40,000-80,000 numbers downplay it I'd suggest. Here are some statistics from Deepseek
Category Region Number of Speakers Source & Year
Some Ability Republic of Ireland 1,873,997 (40% of population) 2022 Census
Some Ability Northern Ireland 228,600 (12.4% of population) 2021 Census
Daily Speakers Republic of Ireland 71,968 2022 Census
Daily Speakers Northern Ireland 43,557 2021 Census
Native/Fluent Global Estimate ~80,000-170,000 Various Sources
Speakers U.S. United States ~20,000+ EstimateRe: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#488Earlier quoted context omitted.
This raises an interesting question. Is there only one dialect of German in the LLM? My understanding is that the German German and Austrian German dialects are significantly different.
My German teacher always claimed that Swiss German and German German (Hochdeutsch) were so different that she needed subtitles to understand it, and she didn't understand why they weren't considered separate languages.
Unless you're thinking of one of the other Swiss languages, Swiss German is actually a variety of Hochdeutsch.
Historically, Germany used to be divided into countless small fiefdoms and each of them used to speak unique barely intelligible languages.
Hochdeutsch is in opposition to Niederdeutsch which Dutch and arguably English are a variety of.
Re: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#489Earlier quoted context omitted.
All the models from all the big providers (even the Chinese models!) support all of these languages already. The big win for accessibility has already been won...3 years ago.
I asked a Finnish person how good an answer about the language example from ChatGPT was. It turned out to be a hallucination, a confidently sounding nonsense. The quality of internet trained models degrade very fast with language material size
I live in Finland, ChatGPT is great at Finnish and far better at it than EuroLLM.
EuroLLM doesn't even beat the open source Gemma 2, and Gemma 3 is already out.
Re: EuroLLM: LLM made in Europe built to support all 24 official EU languages
#490Earlier quoted context omitted.
By your reasoning, English isn't a Germanic language since over half of its vocabulary comes from Latin or French.
You certainly wouldn’t call English a “dialect”
"A language is a dialect with an army and navy" -- Max Weinrich
In the Yiddish original: "אַ שפּראַך איז אַ דיאַלעקט מיט אַן אַרמיי און פֿלאָט", see: https://en.wikipedia.org/wiki/A_language_is_a_dialect_with_a...