comma right after 'and' in lists
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. 19 of Qwen3-0.6B. Its strongest snippets are very consistent: they mostly peak on the same token. 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
The neuron fires on a comma that comes straight after the word 'and' in a list, like 'routine and, mental well-being'. It is pushed the other way in computer code, where it reacts to the equals sign and type names. Steering did not show a clear theme.
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.
-
child's development and, on the 'Words for Life' website, there are some useful tips about how
-
the international humanitarian system and, of course, refugees themselves. A new RAND report details how technology can be part
-
facilitate habits and, for some, sustained compulsivity and addiction. Like other complex illnesses and disorders,
-
culture, history, and society.⏎About the authors⏎Elaine Treharne is Roberta Bowman Denning Professor of Humanities and Professor of English (and, by courtesy, of German Studies) at Stanford University, as well as Director of
-
the other code (too tired to check who it was that posted it, and I can't see just now anyway :P)⏎void genFactorial(int x, int t)⏎int total = t; // you probably wouldn't even need total
-
personal lifestyle of a person includes a daily diet, working conditions, exercise routines and, mental well-being. This examination helps the medical practitioner to identify key symptoms that are
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: there, there, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing, nothing,
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
1.DataSource, DataSet) ' Get the DataTableCollection. Dim tablesCollection As DataTableCollection = thisDataSet.Tables ' Get a specific table by name. Dim table As DataTable = tablesCollection("Suppliers") Console.WriteLine(table.TableName) End Sub⏎Available
-
ster March 3 2012 import random verb = ['erratic', 'unbelievable', 'unfortunate', 'inconsequential'] fruit = ['raspberry', 'pomegranite', 'bananana',
-
to check who it was that posted it, and I can't see just now anyway :P)⏎void genFactorial(int x, int t)⏎int total = t; // you probably wouldn't even need total now⏎void main(void)
-
memchr ( const void *, int, size_t );⏎instead of the two overloaded versions provided in C++.⏎/* memchr example */⏎int main ()⏎char * pch;⏎char str = "Example string";⏎pch = (char
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
- L18:2702 suppresses
- L18:1948 excites
- L13:19 excites
- L12:1163 suppresses
- L17:1946 weakly suppresses
- L5:110 weakly suppresses
- L8:7 weakly excites
- L18:1367 weakly suppresses
- L15:2495 weakly suppresses
- L14:2423 weakly suppresses
- L8:3 weakly excites
- L17:503 weakly suppresses
- L12:2418 weakly suppresses
- L16:2987 weakly suppresses
- L18:2775 weakly excites
- L14:2923 weakly suppresses
- L10:81 weakly suppresses
- L3:456 weakly suppresses
- L18:741 weakly suppresses
- L12:177 weakly excites
- L12:1921 weakly suppresses
- L3:2140 weakly suppresses
- L12:1681 weakly suppresses
- L14:441 weakly excites
Feeds later layers
- L20:1479 strongly suppresses
- L22:539 strongly suppresses
- L22:1710 strongly suppresses
- L20:1768 suppresses
- L24:334 suppresses
- L25:2646 suppresses 'there' after a because or if clause
- L23:434 suppresses
- L22:1300 suppresses
- L20:333 excites
- L23:1295 suppresses
- L20:1985 excites
- L26:2533 suppresses
- L20:218 suppresses
- L20:46 excites
- L24:1318 suppresses
- L21:1871 suppresses
- L20:1873 suppresses
- L25:1229 suppresses
- L20:126 excites
- L25:976 suppresses
- L20:1960 suppresses
- L26:406 suppresses
- L22:1562 weakly suppresses
- L23:67 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
- except
- stär
- ypress
- except
- plementary
- vron
- zeug
- 久了
- ><?=$
- until
Suppresses
- none
- None
- NULL
- null
- None
- null
- nothing
- Nothing
- none
- .NONE