Model card for ClimAdaptLM-I-chunk-classifier

Model description

This model is a fine-tuned version of climatebert/distilroberta-base-climate-f that classifies chunks of climate policy texts as Adaptation or Not adaptation.

It is fine-tuned on a labelled dataset of 3,159 chunks with an average of 3,158 characters and 10 paragraphs per chunk. More details and the methodology can be found in the corresponding research paper and on our Git.

Model performance

Recall: 0.871 F1 score: 0.759 Precision: 0.673

Citation information

@inproceedings{bonen2025transforming,
    title={{Transforming adaptation tracking: Benchmarking Transformer-based NLP approaches to retrieve adaptation-relevant information from climate policy text}},
    author={Bonenkamp, Jetske and Biesbroek, Robbert and Athanasiadis, Ioannis},
    booktitle={Proceedings of The 2nd Workshop of Natural Language Processing meets Climate Change},
    year={2025}
}