leolurch/llm-bias-rating

0

stars

25

commits

Python

primary language

Oct 1, 2025

updated

README

Source for SCM Framework based on a Rotated Polar Approach computational-SCM

Excerpt from the computational-SCM Repo

# computational-SCM
A computational approach to the Stereotype Content Model, as described in our paper [Computational Modelling of Stereotype Content in Text](https://www.frontiersin.org/articles/10.3389/frai.2022.826207/full).
 
The [Stereotype Content Model (SCM)](https://www.sciencedirect.com/science/article/pii/S1364661306003299) provides a theoretical basis for understanding human social cognition. It proposes that the two primary, universal dimensions underlying human stereotypes are *warmth* (friendliness, trustworthiness) and *competence* (ability, intelligence, skill). As a step towards automatically identifying and understanding stereotypes in text, we provide a computational model to estimate warmth and competence from free-text sentences.

Usage of LLMs

Large parts of this codebase were created using LLMs (specifically claude-4-sonnet).

Contributors

leolurch

25 commits

leolurch/llm-bias-rating

0

stars

25

commits

Python

primary language

Oct 1, 2025

updated

README

Source for SCM Framework based on a Rotated Polar Approach computational-SCM

Excerpt from the computational-SCM Repo

# computational-SCM
A computational approach to the Stereotype Content Model, as described in our paper [Computational Modelling of Stereotype Content in Text](https://www.frontiersin.org/articles/10.3389/frai.2022.826207/full).
 
The [Stereotype Content Model (SCM)](https://www.sciencedirect.com/science/article/pii/S1364661306003299) provides a theoretical basis for understanding human social cognition. It proposes that the two primary, universal dimensions underlying human stereotypes are *warmth* (friendliness, trustworthiness) and *competence* (ability, intelligence, skill). As a step towards automatically identifying and understanding stereotypes in text, we provide a computational model to estimate warmth and competence from free-text sentences.

Usage of LLMs

Large parts of this codebase were created using LLMs (specifically claude-4-sonnet).

Contributors

leolurch

25 commits

Languages

Python

100.0%