number and sequence words in math text
A neuronOne of thousands of tiny dials inside each layer. It reads the text so far and produces a number, its activation, that nudges what the model writes next. in layerOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. 6 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featuredA neuron picked for the full treatment: steering outputs and a written explanation. About 3,000 of the 86,000 neurons are featured., so it has steering outputs.
What this neuron does
This neuron tends to light up on words about numbers, sequences and patterns, like 'sequence', 'number' and 'missing numbers'. Turning it up made the model drift toward number puzzles and counting.
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 activationThe number a neuron produces for a token. Large positive means it fired hard; negative means it pushed the other way. was highest. Amber marks positive activation; the outlined tokenA piece of text the model reads at once: often a word, sometimes part of a word or a punctuation mark. is the peak. Hover a token for its strength.
-
0.07 members after 200. As we continue checking larger numbers and see that they do not belong to the sequence, the probability that the sequence contains more elements vanishes
-
85), and Odious Debts: Loose Lending, Corruption and the Third World”s Environmental Legacy
-
asked to look at the grid of numbers and identify the missing numbers and then fill in the missing numbers to count from 1 to 25. In these missing number worksheets some of the numbers are already filled in to help kids to recognize the
-
examples of patterns you could use:⏎- Three letters, special character, three numbers⏎- Three numbers, special character, three letters⏎- Take a word and insert numbers and special characters in specific positions⏎Those are the basic components of your new
-
related to the persistence of chryptophytes and chlorophytes in CM. Multivariate
-
second number is always lower than the first one. No number will begin with zero.⏎For each expression, print two lines with two given numbers
SteeringAdding a push along a neuron while the model writes, to see how its output changes. Turning it up or down.
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.
Here is something I have been thinking about lately: the number of the month of April, the first time the number of the month of April, the first time the number of the month of April, the first time the number of the month of April, the first time the number of the
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
David de Boivil witnessing a charter as early as 1164.<|endoftext|>For the questions below, use the most efficient Excel formula or procedure to find your answer. Calculations by hand (or calculator) should only be used to check
-
writing existing maps in the target location and mapset should be allowed.⏎The next prompt asks you to select one of two windows:⏎Please select one of the following options
-
as a percent in a variety of ways. You can even make use of a mark percentage calculator to calculate your marks. If you’re unsure of how to write a number, try a calculator
-
options contract has the following properties or attributes:⏎- Expiration Date⏎- Symbol or Underlying⏎-
Connections
Show on mapWhich neuronsOne of thousands of tiny dials inside each layer. It reads the text so far and produces a number, its activation, that nudges what the model writes next. in earlier layersOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.
Fed by earlier layers
- L5:62 weakly excites
- L3:703 weakly suppresses
- L1:2524 weakly excites
- L5:60 weakly excites
- L3:52 weakly suppresses
- L2:41 weakly suppresses
- L5:2128 weakly excites
- L3:2752 weakly suppresses
- L2:2709 weakly excites
- L3:278 weakly excites
- L3:731 weakly suppresses
- L3:604 weakly excites
- L2:36 weakly excites
- L2:2845 weakly excites
- L4:2054 weakly suppresses
- L3:1867 weakly suppresses
- L2:2275 weakly suppresses
- L4:101 weakly suppresses
- L5:81 weakly suppresses
- L2:1131 weakly excites
- L4:382 weakly suppresses
- L5:2383 weakly suppresses
- L3:461 weakly excites
- L3:574 weakly suppresses
Feeds later layers
- L9:1230 excites
- L10:1850 excites
- L10:924 excites
- L7:1811 excites
- L13:2793 suppresses
- L9:26 excites
- L19:1531 suppresses
- L13:3009 weakly excites
- L9:1882 weakly excites
- L10:1658 weakly excites
- L15:723 weakly suppresses
- L7:61 weakly suppresses
- L11:2033 weakly excites
- L9:489 weakly excites
- L12:1534 weakly excites
- L12:166 weakly suppresses
- L7:139 weakly suppresses
- L9:2151 weakly excites
- L15:949 weakly excites
- L10:1950 weakly excites
- L17:1005 weakly excites
- L8:2498 weakly suppresses
- L9:627 weakly suppresses
- L7:772 weakly suppresses
Activation histogramA chart of how often a neuron takes each activation value.
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 lensReading which output tokens a neuron pushes toward or away from, straight off its output weights, before the later layers have their say.
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 layersOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. instead.
Promotes
- .EditorButton
- pan
- 花样
- ")[
- [Z
- ToStr
- theast
- .Usage
- inct
- ,"%
Suppresses
- leve
- s
- Rouge
- stage
- adam
- o
- a
- 信任
- си
- 竣工