Edit model card

This model is a fine-tuned version of RoBERTa-large [1]. It was trained on 1,400 employee reviews to measure corporate culture. More specifically, it predicts which of the four culture dimensions of the Competing Values Framework [2,3] best fits the text at hands, i.e., what is the dominant culture.

The model assigns one of four possible labels:

0 (clan): Text is best described by a clan (collaborating) culture.
1 (adhocracy): Text is best described by an adhocracy (creating) culture.
2 (market): Text is best described by a market (competing) culture.
3 (hierarchy): Text is best described by a hierarchy (controlling) culture.

For details on the model and its performance, see Koch and Pasch (2023). Please cite this article when using the model:
S. Koch and S. Pasch, "CultureBERT: Measuring Corporate Culture With Transformer-Based Language Models," in 2023 IEEE International Conference on Big Data (BigData), pp. 3176-3184. doi: 10.1109/BigData59044.2023.10386765

Please see the following tutorial on how to apply CultureBERT to measure corporate culture in your own text documents: https://github.com/Stefan-Pasch/CultureBERT

Other References:

[1] Liu, Y.; Ott, M.; Goyal, N.; Du, J.; Joshi, M.; Chen, D.; ... & Stoyanov, V. (2019). Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.

[2] Cameron, Kim S.; Quinn, Robert E. (2011): Diagnosing and Changing Organizational Culture. Based on the Competing Values Framework. 3rd ed. San Francisco (CA): Jossey-Bass.

[3] Quinn, Robert E.; Rohrbaugh, John (1983): A Spatial Model of Effectiveness Criteria: Towards a Competing Values Approach to Organizational Analysis. In Management Science 29 (3), pp. 363–377. DOI: 10.1287/mnsc.29.3.363.

Downloads last month
15
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.