'there' after a because or if clause

What this neuron does

This neuron lights up on the word 'there' when it starts a clause after words like 'as', 'so' or 'even when', as in 'as there is no force'. Turning it up made the model repeat 'there' over and over, so the link to that word is clear.

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

   learning are predicting customer churn, personalized recommendations, fraud detection, etc.⏎Unsupervised machine learning⏎Unsupervised learning is a machine learning technique, in which there is no need to supervise the model. Instead, the model works on its own

2. #2 peak

   and speed by operating (some of) the controls of the vehicle. A fully automated vehicle does not have a driver in the sense of the Conventions, as there is not a human that decides on the speed and direction (not to be mistaken

3. #3 peak

   and fire then turns that fuel back into carbon dioxide, water and mineral ash. Rain prompts another turn of the great wheel. If there was no plant life, there would be no fire.⏎Viruses are also an expression of cellular life. Vir

4. #4 peak

   On a lifeless, airless satellite there would never be any scavengers or souvenir hunters, no wind to bury or wear down the abandoned spacecraft and artifacts, and no air to corrode metal. Even the footprints would still be

5. #5 peak

   missing is the familiar sensation of weight. Without air resistance, all objects fall at the same rate. Because you and your spacecraft are falling at the same rate, there is no force between you and it to provide a

6. #6 peak

   they started. To reliably control the quantum information, an asymmetric light wave is needed.⏎“The optimum would be a completely directional, unipolar ‘wave’, so there would be only the central peak, no oscillations. That would be the dream

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

   War on Terror’. Curiously, despite their proliferation, they have received relatively little academic attention. In A Theory of the Drone, French philosopher Grégoire Chamayou seeks to comprehend how drones have revolutionised warfare and to deconstruct the

2. #2 peak

   Grégoire Chamayou seeks to comprehend how drones have revolutionised contemporary warfare and to deconstruct

3. #3 peak

   a graphic designer. He designed posters, brochures, album covers, etc. He many times cooperated with other artists, for example György Rad

4. #4 peak

   ik (1898), 32; Lee, The Microtomists Vade Mecum (London, 1900); Macallum, On the Detection and Localization of Phosphorus in Animal and Vegetable Cells

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.