AI interpretability

Biohub ESMFold2 / ESMFold2-Fast all-atom co-folding (Candido et al.

JimLiu/science-skills

Adds a new anomaly-detection model to anomalib under src/anomalib/models/.

open-edge-platform/anomalib

Remove refusal behaviors from open-weight LLMs using OBLITERATUS — mechanistic interpretability techniques (diff-in-means, SVD, whitened SVD, LEACE, SAE decomposition, etc.) to excise guardrails…

RedWoodOG/Hermes-Desktop

Look up what a Biohub ESM-C sparse-autoencoder (SAE) feature means — its label, description, top-activating proteins, decoder neighbours, and activation statistics — by querying the Biohub…

softnanolab/bagel

Guides training and analyzing sparse autoencoders with SAELens to break neural network activations into interpretable features, including superposition and monosemanticity studies.

Orchestra-Research/AI-Research-SKILLs

Guides mechanistic interpretability work with TransformerLens: loading models, caching activations, using HookPoints, activation patching and attention-pattern analysis.

Orchestra-Research/AI-Research-SKILLs

Provides guidance for interpreting and manipulating neural network internals using nnsight with optional NDIF remote execution.

Orchestra-Research/AI-Research-SKILLs

Guides causal experiments on PyTorch models with pyvene, such as causal tracing, activation patching and interchange intervention training, to test how a model works.

Orchestra-Research/AI-Research-SKILLs

Explains machine learning predictions with SHAP: picking the right explainer, computing Shapley values and drawing waterfall, beeswarm, bar and force plots.

davila7/claude-code-templates

Write allocation-efficient buffer code in Corvus.JsonSchema using the codebase's established three-tier pooling pattern: stackalloc → ArrayPool → ThreadStatic caches.

corvus-dotnet/Corvus.JsonSchema

Applies the reasoning style of Geoffrey Hinton, deep learning pioneer and 2018 Turing Award winner.

K-Dense-AI/mimeo

Designing review workflows to surface and mitigate bias in AI outputs.

Owl-Listener/ai-design-skills

Train sparse autoencoders to interpret model features. An agent skill from Luciole-Studio/Misaka-Agent.

Luciole-Studio/Misaka-Agent

Reach for this skill whenever you are discussing reinforcement learning, agentic AI systems, AI alignment, continual learning, or the philosophical limits of large language models.

K-Dense-AI/mimeo

NV-Tesseract Forecasting — transformer-based multivariate time series forecasting with DARR (context-enhanced kNN retrieval), interpretability, and fine-tuning.

NVIDIA/skills

Designing for informed user consent, opt-out, and human override.

Owl-Listener/ai-design-skills

Operational guide for implementing a new Envilder runtime SDK.

macalbert/envilder

Per-feature NaN-safe Spearman/Pearson correlation across many features (genes, proteins, variants) with missing values.

jaechang-hits/SciAgent-Skills

Designs primary, secondary, and exploratory endpoints for biomedical and clinical research protocols.

aipoch/medical-research-skills

Your AI research and engineering brain trust. An agent skill from coco-research/coco.

coco-research/coco

Designs retrospective or prospective clinical cohort study protocols for biomedical and clinical research.

aipoch/medical-research-skills

A skill your agent uses when deciding whether a project is a strong AAAI submission across its broad AI scope, should be reframed or routed to a dedicated track such as AI for Social Impact or AI…

brycewang-stanford/Awesome-Journal-Skills

The creation of effective visualizations is a fundamental component of data analysis.

bioMate-AI/biomate-bioconductor-kb