interpretability

End-to-end pipeline for seeing how LLMs actually process your prompts. Capture attention across every layer, render heatmaps and cooking curves, compare variants with evidence — not vibes.

Tree prompting: easy-to-use scikit-learn interface for improved prompting.

Intriguing Properties of Data Attribution on Diffusion Models (ICLR 2024)

Linear probe found representations of scene attributes in a text-to-image diffusion model

A Multimodal Transformer: Fusing Clinical Notes With Structured EHR Data for Interpretable In-Hospital Mortality Prediction

Interpreting CLIP with Hierarchical Sparse Autoencoders (ICML 2025)

A compiler that transforms computation graphs into the weights of a transformer

Exploring Visual Interpretability for Contrastive Language-Image Pretraining

[🔥 ICLR 2026] - Misaligned Roles, Misplaced Images: Structural Input Perturbations Expose Multimodal Alignment Blind Spots

Reward a Language Model with pancakes 🥞

Tracking exactly what happens to the internal "circuitry" (induction heads) of a 2-layer attention-only Transformer when forced to undergo domain adaptation from prose to structured Python code.