capital N starting a name
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. 27 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
This neuron responds to a capital N that starts a name or word, such as Naperlou, Noreen, Nynke or NRC. It goes negative on the odd lowercase bits that follow a capital letter, like '/N' or '-n'. Turning it up made the model repeat words like 'office' and 'Celsius', and turning it down made the text wander.
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.
-
need to take time and look around. It is amazing what the world has to teach us!⏎Yes, I am also fully agree with you both (Rob and Naperlou). In this case we can take the best example from nature i
-
Anesthetized mice had up to a 25 percent reduction in heart function measures within 20 minutes of exposure to the chemical. Study co-author Nipavan Chiamvimonvat remarked that in this case, triclos
-
a Russian scientist, N.V. Lazarev, who was interested in substances that helped the body adapt
-
components, and for good reason considering all the essential vitamins and nutrients found in milk. Calcium is one of the most well known ones, but according to Dr. Noreen Iftikhar from the healthline newsletter milk also contains magnesium,
-
of billions of dollars, but as the following discoveries prove, the hidden secrets of the seabed are about much more than just money.⏎Billion-dollar wreck: Nuestra Señora de Atocha⏎In July 1985
-
human driving the vehicle, but with the rise of automation, the role of the human driver will slowly disappear. Our author takes a closer look.⏎An article by Nynke E Vellinga⏎Two international road traffic Conventions, namely the
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: Oneoneonononononononononononononononononononononononononononononononononononononononononononon
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
insurance. About 44% of these people were white, 33% were Hispanic, 15% were black, 5% were Asian/Native Hawaiian or Pacific Islander, and 3% were of another race,
-
“Being there for those in difficulty, for refugees and migrants, the poor and the lonely and for the victims and perpetrators of violence and abuse, is a non-negotiable requirement if we are to be faithful to
-
, which give rise to expectations. Frustrated expectations result in an effective response. Music only has meaning unto itself and the music listener ascribes any extra-musical meaning. This includes any emotional meaning. The unfolding of a song is
-
to protect the earth’s climate. The global climate crisis is non-negotiable. #Move the Date: We need to do everything we can to
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
- L22:2299 excites
- L17:391 excites
- L16:55 excites
- L5:499 excites
- L2:1923 suppresses
- L18:127 excites
- L19:595 excites
- L18:282 excites
- L0:210 excites
- L10:431 excites
- L20:212 excites
- L21:2959 excites
- L23:1532 suppresses
- L3:410 suppresses
- L26:1564 excites
- L1:27 excites
- L18:2453 suppresses
- L10:1745 weakly excites
- L16:898 weakly suppresses
- L0:1484 weakly suppresses
- L7:421 weakly suppresses
- L2:238 weakly excites
- L23:206 weakly suppresses
- L11:969 weakly suppresses
Feeds later layers
None: this is the last layer.
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
- ense
- typingsSlinky
- 张
- 发放
- _shuffle
- ChangedEventArgs
- EncodingException
- ----------------------------------------------------------------------
- (records
- 纪
Suppresses
- e
- onn
- on
- lius
- onne
- ll
- ON
- u
- onis
- eos