colon introducing an explanation

What this neuron does

This neuron fires on a colon ":" that comes right after a statement and introduces an explanation, a list or an example. Turning it down made the model fill its text with colons, so the neuron is tied to that punctuation mark.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. Think it is wrong? Flag it (opens a prefilled issue on GitHub).

Fires most on

The text windows where this neuron's activation was highest. Amber marks positive activation; the outlined token is the peak. Hover a token for its strength.

1. #1 peak

   the properties all the solutions a given set of equations have is more important than finding a particular solution: this leads into some of the deepest areas in all of mathematics, both conceptually

2. #2 peak

   in the Miami area who, for whatever reason, cannot get to a⏎safe and dark observing site: check out the locations where the⏎Southern Cross Astronomical Society⏎(Phone

3. #3 peak

   needs of each student.⏎The Referendum funds numerous resources, including technology for media center Maker Spaces, which are collaborative areas where students can create, invent and explore:<|endoftext|>Many

4. #4 peak

   don't want to send all the coins you got from the transaction, so you also specify how much change you are sending to another address, usually controlled by yourself: "I got

5. #5 peak

   the Land Registry proposes the Land Titles Bill by the replacement of current deeds registration system. However, such proposal has no intention to deal with the long neglected land rights: boundary extent.”⏎Hong Kong represents a unique territory where the history of land division,

6. #6 peak

   significant effect to decrease depression score of the mental disorder patients of Banyumas public hospital. Two times cognitive therapy was more effective than once therapy. The cognitive therapy: cognitive restructuritation has a significant effect to decrease dep ression score

Steering

Push this neuron up or down while the model writes, and watch what changes. The outputs were generated ahead of time, so nothing runs in your browser.

Most negative on

Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.

1. #1 peak

   -October, Australia reported over 200,000 cases of the flu. In 2009, they had under 60,000 cases.⏎Even in the United States, where flu season starts later

2. #2 peak

   waged war on White people, especially women.⏎- Asian-Americans are illiterate, undesirable, full of “filth and disease” and unassimilable. They were perceived as “marginal members of the human race,” were denied

3. #3 peak

   so many people now philosophically insist that mandatory mask mandates are unconstitutional? We’ll resist answering that question as it involves a political discussion beyond the scope of this post. But we will textually discuss whether mask mandates are unconstitutional. And they’re not

4. #4 peak

   uth” era where the opinions of experts are viewed with skepticism, where everyone’s opinions are equally valid and opposing views (and evidence) dismissed as “fake news”; and secondly, the sharing of information through the Internet and social media, personalized

Which neurons in earlier layers feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.

Activation histogram

How often this neuron fires at each strength. Almost always it does close to nothing (the tall bars in the middle); the snippets above are the rare moments it fired hard, amber for positive and blue for negative.

Logit lens

The words this neuron nudges the model toward (amber) or away from (blue) when it fires. A hint, not a label: many neurons work through later layers instead.