Datasets:

openeurollm
/

EuroMultiSynth-Instruct

license: apache-2.0
language:
  - bg
  - bs
  - ca
  - cs
  - da
  - de
  - el
  - es
  - et
  - eu
  - fi
  - fr
  - ga
  - gl
  - hr
  - hu
  - is
  - it
  - ka
  - lt
  - lv
  - mk
  - mt
  - nl
  - 'no'
  - pl
  - pt
  - ro
  - sk
  - sl
  - sq
  - sr
  - sv
  - tr
  - uk
task_categories:
  - text-generation
tags:
  - synthetic
  - instruction-tuning
  - sft
  - multilingual
  - llm-as-a-judge
size_categories:
  - 1M<n<10M
configs:
  - config_name: all
    default: true
    data_files:
      - split: train
        path: data/*/train-*.parquet
  - config_name: bg
    data_files:
      - split: train
        path: data/bg/train-*.parquet
  - config_name: bs
    data_files:
      - split: train
        path: data/bs/train-*.parquet
  - config_name: ca
    data_files:
      - split: train
        path: data/ca/train-*.parquet
  - config_name: cs
    data_files:
      - split: train
        path: data/cs/train-*.parquet
  - config_name: da
    data_files:
      - split: train
        path: data/da/train-*.parquet
  - config_name: de
    data_files:
      - split: train
        path: data/de/train-*.parquet
  - config_name: el
    data_files:
      - split: train
        path: data/el/train-*.parquet
  - config_name: es
    data_files:
      - split: train
        path: data/es/train-*.parquet
  - config_name: et
    data_files:
      - split: train
        path: data/et/train-*.parquet
  - config_name: eu
    data_files:
      - split: train
        path: data/eu/train-*.parquet
  - config_name: fi
    data_files:
      - split: train
        path: data/fi/train-*.parquet
  - config_name: fr
    data_files:
      - split: train
        path: data/fr/train-*.parquet
  - config_name: ga
    data_files:
      - split: train
        path: data/ga/train-*.parquet
  - config_name: gl
    data_files:
      - split: train
        path: data/gl/train-*.parquet
  - config_name: hr
    data_files:
      - split: train
        path: data/hr/train-*.parquet
  - config_name: hu
    data_files:
      - split: train
        path: data/hu/train-*.parquet
  - config_name: is
    data_files:
      - split: train
        path: data/is/train-*.parquet
  - config_name: it
    data_files:
      - split: train
        path: data/it/train-*.parquet
  - config_name: ka
    data_files:
      - split: train
        path: data/ka/train-*.parquet
  - config_name: lt
    data_files:
      - split: train
        path: data/lt/train-*.parquet
  - config_name: lv
    data_files:
      - split: train
        path: data/lv/train-*.parquet
  - config_name: mk
    data_files:
      - split: train
        path: data/mk/train-*.parquet
  - config_name: mt
    data_files:
      - split: train
        path: data/mt/train-*.parquet
  - config_name: nl
    data_files:
      - split: train
        path: data/nl/train-*.parquet
  - config_name: 'no'
    data_files:
      - split: train
        path: data/no/train-*.parquet
  - config_name: pl
    data_files:
      - split: train
        path: data/pl/train-*.parquet
  - config_name: pt
    data_files:
      - split: train
        path: data/pt/train-*.parquet
  - config_name: ro
    data_files:
      - split: train
        path: data/ro/train-*.parquet
  - config_name: sk
    data_files:
      - split: train
        path: data/sk/train-*.parquet
  - config_name: sl
    data_files:
      - split: train
        path: data/sl/train-*.parquet
  - config_name: sq
    data_files:
      - split: train
        path: data/sq/train-*.parquet
  - config_name: sr
    data_files:
      - split: train
        path: data/sr/train-*.parquet
  - config_name: sv
    data_files:
      - split: train
        path: data/sv/train-*.parquet
  - config_name: tr
    data_files:
      - split: train
        path: data/tr/train-*.parquet
  - config_name: uk
    data_files:
      - split: train
        path: data/uk/train-*.parquet

EuroMultiSynth-Instruct: synthetic instruction-tuning data in 35 European languages

4,160,598 single-turn instruction/response pairs in 35 European (and neighbouring) languages, written natively in each language (not translated) by an LLM, scored by an LLM judge, quality-filtered and de-duplicated within each language. Built with synthgen for the OpenEuroLLM project.

Quick start

from datasets import load_dataset

ds = load_dataset("openeurollm/EuroMultiSynth-Instruct", "de", split="train")   # one language
ds = load_dataset("openeurollm/EuroMultiSynth-Instruct", split="train")         # all 35 languages
print(ds[0]["messages"])

Fields

field	type	description
`id`	string	unique row id, `<lang>-<number>`
`language`	string	ISO 639-1 code
`messages`	list of `{role, content}`	`[user prompt, assistant response]`
`domain`	string	task category the prompt was generated for (18 values)
`persona`	string or null	user persona the prompt was written as (null for 87% of rows)
`judge_score`	int8	LLM-judge quality score, 0-10 (only 7-10 kept)

How it was made

Generation. google/gemma-4-31b-it (served with vLLM) wrote each instruction and its response in the target language in one call, conditioned on a task domain and, for some prompts, a user persona. 7,040,000 unique samples were generated (160k or 320k per language).
Judging. Qwen/Qwen3.6-27B scored the samples 0-10 (temperature 0, thinking disabled). 2,204 samples (0.03%) were never judged and are not included.
Quality filter. Kept samples with score ≥ 7: 5,688,783 rows (80.9% of scored samples).
Exact de-duplication on (language, whitespace/case-normalised prompt): −93,387 rows.
Near-duplicate removal, within each language only. Prompts were embedded and a prompt was dropped if its cosine similarity to an earlier kept prompt in the same language was at or above the threshold (greedy, first occurrence kept). 31 languages used sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 at 0.97; Maltese, Irish, Basque and Icelandic, which that model does not cover, used sentence-transformers/LaBSE at 0.96 (calibrated on German and Croatian to remove the same share as MiniLM at 0.97). Thresholds were chosen from sweeps with example pairs per similarity band: at these values, removed pairs are rewordings of the same request. −1,434,798 rows.

Final: 4,160,598 rows.

Languages

code	language	rows	near-dup model / threshold
`tr`	Turkish	236,352	MiniLM ≥ 0.97
`de`	German	230,459	MiniLM ≥ 0.97
`bg`	Bulgarian	212,388	MiniLM ≥ 0.97
`bs`	Bosnian	207,040	MiniLM ≥ 0.97
`gl`	Galician	193,629	MiniLM ≥ 0.97
`hr`	Croatian	190,139	MiniLM ≥ 0.97
`da`	Danish	185,955	MiniLM ≥ 0.97
`ca`	Catalan	166,246	MiniLM ≥ 0.97
`fi`	Finnish	132,641	MiniLM ≥ 0.97
`pl`	Polish	126,767	MiniLM ≥ 0.97
`it`	Italian	124,420	MiniLM ≥ 0.97
`ro`	Romanian	119,021	MiniLM ≥ 0.97
`no`	Norwegian	115,400	MiniLM ≥ 0.97
`mk`	Macedonian	110,730	MiniLM ≥ 0.97
`sr`	Serbian	110,270	MiniLM ≥ 0.97
`uk`	Ukrainian	109,240	MiniLM ≥ 0.97
`sq`	Albanian	107,505	MiniLM ≥ 0.97
`sk`	Slovak	106,467	MiniLM ≥ 0.97
`sv`	Swedish	104,862	MiniLM ≥ 0.97
`pt`	Portuguese	104,727	MiniLM ≥ 0.97
`fr`	French	103,013	MiniLM ≥ 0.97
`is`	Icelandic	102,362	LaBSE ≥ 0.96
`nl`	Dutch	102,213	MiniLM ≥ 0.97
`cs`	Czech	100,099	MiniLM ≥ 0.97
`es`	Spanish	99,709	MiniLM ≥ 0.97
`el`	Greek	99,005	MiniLM ≥ 0.97
`ka`	Georgian	90,665	MiniLM ≥ 0.97
`sl`	Slovenian	85,119	MiniLM ≥ 0.97
`lt`	Lithuanian	71,445	MiniLM ≥ 0.97
`hu`	Hungarian	68,390	MiniLM ≥ 0.97
`lv`	Latvian	62,438	MiniLM ≥ 0.97
`mt`	Maltese	52,674	LaBSE ≥ 0.96
`et`	Estonian	52,551	MiniLM ≥ 0.97
`eu`	Basque	50,716	LaBSE ≥ 0.96
`ga`	Irish	25,941	LaBSE ≥ 0.96

Domains

domain	rows	share
general question answering	855,672	20.6%
coding	585,753	14.1%
brainstorming & idea generation	555,206	13.3%
reports, summaries & documentation	287,952	6.9%
roleplay & character play	223,106	5.4%
translation	195,296	4.7%
math (word problems & explanations)	179,613	4.3%
short stories & fiction	165,221	4.0%
social media & blog posts	136,750	3.3%
data analysis (tables, lists, datasets)	133,886	3.2%
summarization	124,134	3.0%
meetings, planning & project coordination	108,337	2.6%
workplace email & communication	104,977	2.5%
humor, jokes & satire	104,715	2.5%
poetry & song lyrics	104,456	2.5%
customer & client communication	99,965	2.4%
job applications & CVs	99,951	2.4%
personal essays & reflections	95,608	2.3%

Judge scores

score	rows	share
7	283,975	6.8%
8	168,940	4.1%
9	476,087	11.4%
10	3,231,596	77.7%

Personas (8): office worker (69,697), teenager (69,427), retiree (69,304), small-business owner (68,846), recent immigrant (68,527), parent (68,376), university student (67,856), tourist (63,608).

Limitations

Fully synthetic. Both the prompts and the responses are model-generated; facts, figures and local details (laws, prices, places) can be wrong or out of date.
LLM-judged. Quality filtering relies on a single LLM judge, whose reliability likely varies by language and is probably weaker for lower-resource languages.
Uneven size and quality across languages. Languages with lower judge pass rates are smaller (e.g. Irish: 320,000 generated, 25,941 kept). A few spot-checked Basque and Irish samples contained malformed words, so expect more noise in the lower-resource languages.
Residual redundancy. The near-duplicate thresholds are deliberately conservative, so many prompts still share a template or scenario with different specifics.
Single-turn only.

License

Released under the Apache License 2.0.

This choice follows from the models used to build the dataset, all of which are Apache-2.0:

role	model	license
generation (wrote every prompt and response)	`google/gemma-4-31B-it`	Apache-2.0
quality judge (scores only, no text in the dataset)	`Qwen/Qwen3.6-27B`	Apache-2.0
near-duplicate filtering	`sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2`, `sentence-transformers/LaBSE`	Apache-2.0

All text in this dataset is model output, attributed to google/gemma-4-31B-it. Please credit both the dataset and the generating model when you use it, and follow Google's Gemma Prohibited Use Policy in spirit when building on Gemma-generated data.

Citation

TODO