search

msakarvadia / llm_bias

Investigating if we can find circuits in LLMs that reinforce human-biases found in training data
MIT License
0 stars 0 forks source link

Clustering #4

Open msakarvadia opened 4 months ago

msakarvadia commented 4 months ago

Try clustering after filtering for the "relavent category" and assess if we can see nice clusters.