---
title: "Mechanistic Interpretability"
description: "Mechanistic interpretability is the study of how neural networks compute their outputs by reverse-engineering internal mechanisms such as features, circuits, attention heads, activations, and weights. It aims to explain what a trained model is doing internally, not only whether its output is correct."
canonical_url: "https://promptmetheus.com/resources/llm-knowledge-base/mechanistic-interpretability"
---

> Index of every page on this site: [llms.txt](https://promptmetheus.com/llms.txt)

Mechanistic interpretability is the study of how neural networks compute their outputs by reverse-engineering internal mechanisms such as features, circuits, attention heads, activations, and weights. It aims to explain what a trained model is doing internally, not only whether its output is correct.

In [Large Language Models (LLMs)](https://promptmetheus.com/resources/llm-knowledge-base/large-language-model-llm), mechanistic interpretability research often investigates:

- which internal features correspond to concepts or behaviors;
- how attention heads move information through the residual stream;
- how circuits implement tasks such as copying, induction, or refusal;
- whether a feature causally affects model output; and
- how model internals change under fine-tuning or steering.

Mechanistic interpretability is different from ordinary evaluation. An [evaluation](https://promptmetheus.com/resources/llm-knowledge-base/evaluation) measures external behavior. Mechanistic interpretability tries to identify internal computations that produce that behavior.

The field is technically difficult because modern models have billions or trillions of [model parameters](https://promptmetheus.com/resources/llm-knowledge-base/model-parameter), distributed representations, and nonlinear interactions across layers. Results that explain a small circuit may not generalize to all model behavior.

Anthropic's [Transformer Circuits Thread](https://transformer-circuits.pub/) collects research on reverse-engineering transformer internals and related interpretability methods.

## Sitemap

See the full [sitemap](https://promptmetheus.com/sitemap.md) for all pages.