Dataset · rlacombe
Climate claims expert evaluation dataset
rlacombe/ClimateX
8,094 IPCC report statements with expert annotation for evaluating LLM climate knowledge.
What's inside
The dataset contains 8,094 statements extracted or paraphrased from the three working group reports of IPCC Assessment Report 6. Each statement is annotated by experts for classification and feature extraction tasks. The data is in English, with a volume between 1K and 10K records.
What you can do with it
- Evaluating large language models' knowledge in climatology
- Training classifiers for expert confidence in scientific texts
- Tasks
- Zero shot classification, Text classification, Embeddings
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 189
- Likes
- 8
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("rlacombe/ClimateX")