the very first token '0'

What this neuron does

This neuron fires strongly on the number 0 when it is the very first thing in a piece of text, and the examples all look alike for that reason. Its other activations are weak and unclear. Turning it up or down produced garbled, code-like or repeated words, so it does not show a clear meaning.

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

   0 largest GHG emitting industrial facilities reduced emissions by 5.6 MTCO

2. #2 peak

   0⏎network => 172.16.0.0/1

3. #3 peak

   0 devices 2^n ~= 25000 n = 15

4. #4 peak

   0-2011, the average number of icebergs to pass below

5. #5 peak

   017).⏎Approximately 65,000 individuals in the U

6. #6 peak

   013), USA Basketball has designed a practical, functional and sequential development model to

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

   iness, breathing difficulties and vomiting.⏎In some cases, poisoning may result in marked psychological

2. #2 peak

   Diff Views⏎By default, GitKraken’s Git diff tool will display in H

3. #3 peak

   angled away from the Sun is still cast in dark shadow, but⏎because of the

4. #4 peak

   abnormal cells that grows out of control.⏎What is cancer? – Video⏎A short

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.