Dataset · ikhlasulakmalh
Multilingual model responses for LLM uniformity research
ikhlasulakmalh/multilingual-hivemind-responses
From 100k to 1M generated language model responses to Infinity-Chat100 prompts in multiple languages.
What's inside
The dataset contains tabular and text data from the multilingual-artificial-hivemind project. It includes language model responses to prompts from the Infinity-Chat100 set across several studied languages. For each model and prompt, exactly 20 generation variants are provided (with temperature 1.0, top-p 0.9, and a 2048 token limit without reasoning mode).
What you can do with it
- Studying the repeatability and diversity of LLM responses across languages
- Comparing generative patterns of various language models
- Size
- 1M–10M rows
- Downloads this month
- 24.1K
- Likes
- 0
- Access
- on request on Hugging Face
How to load it
from datasets import load_dataset
ds = load_dataset("ikhlasulakmalh/multilingual-hivemind-responses")