Mitigating positional bias in LLaVA 1.5 (7B) on ScienceQA via per-head activation steering & PCA.
-
Updated
May 17, 2025 - Jupyter Notebook
#
scienceqa
Here are 2 public repositories matching this topic...
Fine tunned PaliGemma vision-language models using the ScienceQA dataset for visual question answering.
-
Updated
Oct 23, 2024 - Jupyter Notebook
Improve this page
Add a description, image, and links to the scienceqa topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the scienceqa topic, visit your repo's landing page and select "manage topics."