Post-hoc Interpretability of LDA on Hyperspectral Imagery: SHAP Attributions, Counterfactual Topic Flips, and LLM-judge Alignment under Token-Mass-Dispersion Asymmetry

Post-hoc Interpretability of LDA on Hyperspectral Imagery: SHAP Attributions, Counterfactual Topic Flips, and LLM-judge Alignment under Token-Mass-Dispersion Asymmetry --- Source: OpenAlex - Zenodo (CERN European Organization for Nuclear Research) (Citations: 25) PDF: https://doi.org/10.5281/zenodo.21504113 Original Link: https://doi.org/https://doi.org/10.5281/zenodo.21504113