Weight space learning experiments (interpreting behavior through activation signatures)
Maximus Powers
maximuspowers
AI & ML interests
Data science, information theory, philosophy of mind
Recent Activity
updated
a dataset
1 day ago
maximuspowers/muat-pca-10-medium
updated
a dataset
1 day ago
maximuspowers/muat-mean-std-large
published
a dataset
1 day ago
maximuspowers/muat-mean-std-large
Organizations
A study benchmarking major LLM's internalized philosophies, results can be found here: https://philosophai-results.vercel.app/ (view 07-2025 data)
-
maximuspowers/bert-philosophy-classifier
Text Classification • 0.1B • Updated • 9 -
maximuspowers/philosophai-mcq-benchmark
Viewer • Updated • 100 • 9 -
maximuspowers/philosophy-schools-multilabel
Viewer • Updated • 5.67k • 17 -
maximuspowers/philpapers-papers-summarized-labeled
Viewer • Updated • 3.78k • 28 • 1
Social bias detection resources used to create a multi-label token classification model for: Generalizations, Unfairness, and Stereotypes
-
Gusnet V1 Demo
⚡1Identify biased entities in text
-
GUS-Net: Social Bias Classification in Text with Generalizations, Unfairness, and Stereotypes
Paper • 2410.08388 • Published • 2 -
ethical-spectacle/social-bias-ner
Token Classification • Updated • 50 • 3 -
ethical-spectacle/gus-dataset-v1
Viewer • Updated • 3.74k • 51
Weight space learning experiments (interpreting behavior through activation signatures)
Social bias detection resources used to create a multi-label token classification model for: Generalizations, Unfairness, and Stereotypes
-
Gusnet V1 Demo
⚡1Identify biased entities in text
-
GUS-Net: Social Bias Classification in Text with Generalizations, Unfairness, and Stereotypes
Paper • 2410.08388 • Published • 2 -
ethical-spectacle/social-bias-ner
Token Classification • Updated • 50 • 3 -
ethical-spectacle/gus-dataset-v1
Viewer • Updated • 3.74k • 51
A study benchmarking major LLM's internalized philosophies, results can be found here: https://philosophai-results.vercel.app/ (view 07-2025 data)
-
maximuspowers/bert-philosophy-classifier
Text Classification • 0.1B • Updated • 9 -
maximuspowers/philosophai-mcq-benchmark
Viewer • Updated • 100 • 9 -
maximuspowers/philosophy-schools-multilabel
Viewer • Updated • 5.67k • 17 -
maximuspowers/philpapers-papers-summarized-labeled
Viewer • Updated • 3.78k • 28 • 1