Open erikerlandson opened 3 months ago
Experiments for understanding and visualizing the "coverage" of LLM benchmark data in semantic spaces.
Currently prototyping ideas in this notebook: https://github.com/erikerlandson/datascience-wg/blob/main/hack-scripts/model-embed-poc.ipynb
Experiments for understanding and visualizing the "coverage" of LLM benchmark data in semantic spaces.
Currently prototyping ideas in this notebook: https://github.com/erikerlandson/datascience-wg/blob/main/hack-scripts/model-embed-poc.ipynb