Dataset · Text generation
comparia-fr-arena
by Ministère de la Culture (SNUM) ministere-culture/comparia-fr-arena
Compar:IA is a public chatbot arena run by the French Ministry of Culture. People chat with two anonymous models side by side and say which answer they prefer. This dataset is the result.
Dataset Card
Compar:IA is a public chatbot arena run by the French Ministry of Culture. People chat with two anonymous models side by side and say which answer they prefer. This dataset is the result. Each row is one turn of a conversation: the two models' answers to the same user message, the preference the user gave on that turn (if any), and the full conversation both answers belong to. The prompts come from real users, most of them writing in French. With 675k response pairs and 208k human preferences over 115+ models, it is one of the larger French preference datasets available for training and evaluating chat models. A user sends a prompt and gets back two answers from two anonymous models. The…
Excerpt from the card by Ministère de la Culture (SNUM), licensed etalab-2.0.
Details
- Repository
- ministere-culture/comparia-fr-arena
- Publisher
- Ministère de la Culture (SNUM)
- Task category
- Text generation
- Tags
- chatbot-arena, human-feedback, preferences
- Size category
- 100K<n<1M
- Languages
- fr
- Revision
- 3c59a5e926c5f8313ea29de413394782f4b5685a
- Last updated
- 2026-10-17
Files
7 files, 9.4 GB in total.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| comparia-fr-arena.parquet | Data | 9.3 GB | — |
| comparia-fr-arena_samples.jsonl | Data | 39.4 MB | — |
| comparia-fr-arena_samples.tsv | Data | 36.2 MB | — |
| models.json | Data | 178.6 KB | — |
| vote_tags.json | Data | 870 B | — |
| README.md | Documentation | 14.5 KB | — |
| .gitattributes | Repository | 2.6 KB | — |
License and Download
- License
- etalab-2.0
- Access
- Access requested at publisher
Ministère de la Culture (SNUM) grants access through its official repository on Hugging Face.