animals, plants and natural hazards

What this neuron does

This neuron fires on nouns for living things and natural hazards, such as tornadoes, moths, bees, trees, crayfish and sand. It goes quiet or slightly negative on words that end definitions and examples, like 'that', commas and full stops. Turning it up did not produce a clear nature theme in the sample outputs.

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

   present their research findings on hazard preparedness and mitigation. The research teams, which receive some of their funding from the National Science Foundation, study disasters as varied as tornadoes, hurricanes, oil spills, earthquakes and landslides. They showcased projects

2. #2 peak

   and most are comparatively benign. Have some introduced species caused changes that most everyone would agree have been very harmful? Absolutely. By killing timber trees, gypsy moths and the emerald ash borer have caused, and continue to cause, enormous

3. #3 peak

   if you have a history of allergies to insect bites. Some insect bites or stings can provoke very serious allergic reactions (most of the time these are insects like bees or wasps), and nausea may be

4. #4 peak

   As it moves around us, its viewpoint is changing, and that motion is reflected in the asteroid. Try this: look at some distant object like a mountain or tree, hold your thumb at arm’s length, close one eye, and then move

5. #5 peak

   ,000 other forest products. Trees are renewable, biodegradable and recyclable. Wood by-products become such products as vitamins, plastics, vanilla flavoring, photographic film, toothpaste and medicines.⏎By planting 20 million trees

6. #6 peak

   that electric eels use their energy charge in a number of ways, and not just for stunning their prey. When tackling a larger prey, like a large crayfish, the eel will curl

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

   .⏎- Onus Item — Definition,⏎- A payment that is deposited at the same financial institution in which it was drawn or written. For example, a cheque that is written by one bank customer and deposited by

2. #2 peak

   .⏎Container. Containers are the elements that “hold” the source. For example, if a television episode is watched on Netflix, Netflix is the container.⏎Number and volume. Note that the abbreviations vol.

3. #3 peak

   its effect through logical demonstration. (157)⏎a "system" of signs whose elements receive meaning through and within this system. Morse code is an example. (159)⏎superficial and mistaken definition of rhetoric, as a

4. #4 peak

   WikiPedia often edit and re-upload images used to illustrate articles on that site. But in another way, it may not. Cropping or resizing a photograph is different from using the language of graphic design to convey a message.⏎I'm a

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.