Provides guidance for interpreting and manipulating neural network internals using nnsight with optional NDIF remote execution. Use when needing to run interpretability experiments on massive models (70B+) without local GPU resources, or when working with any PyTorch architecture.
8.1
Rating
0
Installs
Machine Learning
Category
Excellent skill for neural network interpretability with nnsight. The description clearly identifies when to use the skill (remote execution on 70B+ models, PyTorch interpretability). Task knowledge is comprehensive with 5 well-structured workflows covering activation analysis, patching, remote execution, cross-prompt sharing, and gradients. Code examples are complete and practical. Structure is clear with logical progression from concepts to workflows to troubleshooting, plus organized reference files. Novelty is strong: remote execution on massive models is genuinely difficult for CLI agents to implement from scratch and would consume many tokens. Minor improvements possible: could be more concise in places, and the comparison table could be positioned earlier. Overall, this is a high-quality skill that meaningfully reduces complexity for interpretability tasks.
davila7
Skill Author
Loading SKILL.md…
