"of" after process or action nouns

What this neuron does

The neuron lights up on "of" when it follows a noun about an action or process, such as "production of", "copy of" or "destruction of". It mostly ignores the word that comes after. Turning it up or down changed the topics the model wrote about but did not show a clear shared theme.

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

   so. Animators spend years studying eye movements to help them create characters with a range of believable expressions, but the eyes of robots and interactive toys are far more difficult to manipulate than those of cartoon characters.⏎

2. #2 peak

   , engines, or generators to control and adjust production of biofuels or biomass-fueled power.⏎- Operate equipment to start

3. #3 peak

   work at height; hazardous substances (COSHH); manual handling; noise; vibration and lead.⏎On construction projects, this may relate not only to the construction of the works, but also to temporary works and to the operation, maintenance, cleaning

4. #4 peak

   / base, oxidant / reducer etc ...) depending on the conditions in which are found in temperature, humidity, and pH, being able to reach levels of destruction of up to 70%.⏎The transit of nutrients through the digestive tract due

5. #5 peak

   1649. They are considered the progenitors of the modern Ukrainian nation and Ukrainian nationalist discourse appeals, to a large extent, to the Cossacks.⏎In 1670, the Don Cossack Stenka

6. #6 peak

   copies are made at least four different ways when accessing the Internet. One way copies are made is by simply viewing a page on the Internet. This causes a copy of that page to be made and stored in the Random Access Memory (RAM) of

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

   8, many activities have been adapted for students ranging from first grade to high school.⏎Number Of Pages:⏎- ID: 97808735525

2. #2 peak

   Reader: A Developmental Approach to Reading Instruction (3rd Edition)⏎|Author||Michael P. O'Donnell, Margo Wood|⏎|Number Of Pages||408|⏎Write a review⏎Note: HTML is not translated!⏎

3. #3 peak

   |The Physical Object|⏎|Number of Pages||136|⏎Learn why the Common Core is important for your child.

4. #4 peak

   common to see creatures drawn with far too many limbs.⏎There are a couple of theories behind this. Some suggest that this was simply because the artists had no means of erasing the limbs, leaving their mistakes on the walls for posterity. Others

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.