nnsight-basics
OfficialTrace, save, and intervene on model internals.
Data & Analytics#interventions#interpretability#model-debugging#nnsight#activation-tracing#gradient-access
Authorndif-team
Version1.0.0
Installs0
System Documentation
What problem does it solve?
NNsight Basics enables researchers and engineers to access and manipulate neural network internals during forward passes, accelerating debugging, hypothesis testing, and targeted interventions.
Core Features & Use Cases
- Access and save intermediate activations across transformer layers during forward passes.
- Intervene on activations (zero out, add noise, replace values) for controlled experiments.
- Support for gradient access, generation workflows, and multi-prompt tracing to study information flow.
Quick Start
Install the nnsight package and start a tracing session to inspect layer outputs and saved tensors.
Dependency Matrix
Required Modules
None requiredComponents
Standard package💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: nnsight-basics Download link: https://github.com/ndif-team/skills/archive/main.zip#nnsight-basics Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.