# Tracing causal structure in LLM-generated text: a different lens on the Dallas circuit

> Source: <https://www.lesswrong.com/posts/uBeXWTAHbyf68yu36/tracing-causal-structure-in-llm-generated-text-a-different>
> Published: 2026-07-20 14:22:09+00:00

The classic "[Dallas](https://transformer-circuits.pub/2025/attribution-graphs/biology.html#dives-tracing)" example from Anthropic focuses on an **internal circuit** in an LLM.

I became curious about what the same underlying process looks like when viewed through the **generated reasoning trace **instead of hidden activations. The resulting attribution graph looks much more structured than I expected:

Below is an animated version:

Using simple gradient attribution, DAG tracing, and sparse pruning on Qwen3-1.7B, the resulting graph already resembles a '**reasoning trajectory**'.

This feels like a different perspective on mechanistic interpretability: instead of analyzing internal circuits directly, it explores causal structure within the generated language itself through context-attribution DAGs. *P.S. There is some related work, such as the ***Thought Anchors*** series. However, the goals are different, and their approach relies on prior semantic knowledge, whereas this approach requires almost no semantic knowledge.*

I'm curious whether this perspective can become useful beyond visualization—for example, as a way to study how information propagates through language during inference, and perhaps, more broadly, the informational dynamics of language and LLMs.

*Source code:* [https://github.com/dqy08/InfoLens](https://github.com/dqy08/InfoLens)