the word 'just' after have or has

What this neuron does

This neuron lights up on the word "just" when it means "a moment ago", as in "has just arrived" or "had just given birth". When it was turned up, the model's writing got stuck repeating "right" in a loop.

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

   it settled slowly behind the northern horizon. During the sunset's final minute, this remarkable sequence of 7 images follows the distorted edge of the solar disk as it just disappears against a distant tree line, capturing both a green and blue flash. Not

2. #2 peak

   goal was to study 40 individual animals over that time in four different categories. The specimens ranged from the plumpest adults to haggard mothers who had just given birth and from doe-eyed pups to those skinny juveniles.⏎Mellish

3. #3 peak

   half minutes before the rover landed. In the photo, the heat shield is roughly 50 feet (16 meters) away from the spacecraft, and had just been separated about three seconds earlier.⏎The heat shield is 15 feet (

4. #4 peak

   RSP<|endoftext|>the story? . . . the Morning Glory cloud rolls majestically above Burketown.⏎The season for Morning Glory cloud⏎formations has just begun in the Gulf of Carpentaria and glider pilots and⏎scientists

5. #5 peak

   Rome from an extended stay in Naples, he wrote that he felt as if "he had just had a cataract removed." For the rest of his stay in Italy,

6. #6 peak

   in the pancreas by little clusters of cells, known as beta cells. Beta cells sense when there is an excess of glucose in the blood stream, such as just after a meal, and they send insulin

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

   by the National Water-Quality Assessment (NAWQA) Program of the U.S. Geological Survey (USGS).1 Chloroform and other trihalomethanes, the most commonly detected compounds, were found in about 9 percent

2. #2 peak

   , Leptoseris cucullata, and less common Madracis formosa, M. pharensis, M. decactis, Montastraea cavernosa, Porites divaricata, Scolymia cub

3. #3 peak

   are formed when organic carbon reacts with chlorine, a chemical used to disinfect water against pathogenic organisms. DBPs include the groups of chemicals known as trihalomethanes and haloacetic acids. Health effects from long-term or chronic exposure

4. #4 peak

   California, Kristin Belair picks the best grapes from 50 acres of vineyards to create high-end cabernet sauvignon and sauvignon blanc wines. Both are part of a growing number

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.