Did changing the context change the answer?

[chenxiachan]

ThoughtDAG is built around a narrow claim: a wire changes what the model receives. I would like to collect concrete cases where that control mattered, including cases where it did not.

If you have one, share a small before-and-after example:

• the question you asked;
• the context change you made, such as removing an edge, choosing a different branch, merging sources, or condensing a run;
• what changed in the answer;
• whether the result was expected or surprising.

Screenshots, exported JSON, or a short written description are all welcome. Please redact private text, API details, file paths, and document content before posting.

Useful context, if you are comfortable sharing it:

• ThoughtDAG version;
• model and approximate context size;
• local or remote inference;
• whether the same prompt was regenerated more than once.

Negative results are especially useful. If removing a branch made no meaningful difference, that is evidence too.