colon introducing a summary point

What this neuron does

This neuron fires on a colon that introduces the main point or example after a sentence like "one thing in common:" or "a surprising discovery:". Turning it up made the text wander into unrelated topics, so the steering doesn't show the same idea.

Written by Claude from this neuron's activations and steering outputs. Confidence: medium. 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

   working women everywhere. The figures represent women of different races and celebrate the contributions made by women throughout time.⏎“Women all over the world have one trait in common: they work, they work hard, and their work is unheralded.”⏎

2. #2 peak

   on the cover of the print edition.⏎Dysprosium is not only the atomic element with the strongest magnetic moments, but it also possesses another interesting property: its magnetic moments point either all the same direction (ferromagnetism) or

3. #3 peak

   out slowly near the trunk, but not on<|endoftext|>Electric cars may be energy efficient Earth-saving futuremobiles, but their rise could also have a strange unintended consequence: killing off the oldest form of radio.⏎Two popular electric cars, the BMW i

4. #4 peak

   using an online image generator.⏎This wasn’t the only piece of misinformation shared during and immediately after the terrorist attack. However, the majority had one thing in common: a visual element. Rarely were stories shared exclusively in text

5. #5 peak

   , scientists have made a surprising discovery: a gigantic aquifer of relatively fresh water trapped in porous sediments lying below the

6. #6 peak

   10 metres).⏎Aside from a few small streams and artificial lakes, Johannesburg lacks water. The city owes its location to the presence of an even more precious resource: gold. The city grew on the edge of the Witwatersrand Main Reef,

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

   strap on our virtual reality goggles, and hit the weights.⏎Researchers from the University of Wisconsin-Madison are doing something like that (although, of course, the weights don't weigh anything) by creating

2. #2 peak

   a common food source for many Native American tribes including but not limited to the Iroquois, Algonquin, Ojibwa, and

3. #3 peak

   of Cardiology: FDA Releases Final Regulation Extending Authority Over E-Cigarettes, Other Tobacco Products<|endoftext|>The aim of the project - which is in the style of a Wikipedia - is to make the large volume of

4. #4 peak

   in the State of Ju”. This story was consolidated into this 4-character idiom and has the original meaning of not forgetting the humble past and in a place like China where the history is considered the backbone of its civilization (all through Chinese history

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.