"the" at the start of a fragment
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. 0 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 most to the word "the" when it is the very first word of a piece of text, which is often where the sample was cut. The negative side fires on "way" in the phrase "along the way." Steering shows no clear theme.
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 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.
-
the Department of Construction, Chaoyang University of Technology, Taichung 4
-
the current ISO Guides.⏎- Basic concepts of measurement⏎- Introduction to the topic measurement
-
the minute ‘energy factories’ within cells.⏎The mitochondria in the mutant mice don
-
the Census then by practically everyone else, themselves included.⏎Census-takers designated them
-
the pleural space. As fluid drains into a collection bottle, you may cough a
-
the end of the month. Some 40 million people live near their shores,
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: I have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately, and
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
.⏎The team is also developing new technology along the way. An effective cooling system does not exist for an E-1 class electric vehicle
-
as smoking.”⏎Note: We’re about to get science-heavy. You can scroll to the bottom for the takeaways. We’ve also summarized important points along the way.⏎The study itself was part epidemiological and part rodent intervention, and can
-
the information by acquiring and practicing these essential auditory skills. The lyrics and music set the stage for listening, learning, and a whole lot of meaningful fun along the way.⏎Take a look at the Read, Sing, Learn: Homophones series
-
a disorder characterized by widespread musculoskeletal pain accompanied by fatigue, sleep and mood issues. Researchers believe that fibromyalgia amplifies painful sensations by affecting the way your brain and spinal cord process painful and nonpainful signals.⏎Fibrom
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
None: this is the first layer.
Feeds later layers
- L5:493 excites
- L10:111 weakly suppresses
- L6:1530 weakly suppresses
- L20:2469 weakly suppresses
- L5:2733 weakly suppresses
- L5:420 weakly suppresses
- L15:1544 weakly excites
- L12:1340 weakly suppresses
- L7:566 weakly suppresses
- L17:727 weakly suppresses
- L9:662 weakly suppresses
- L1:206 weakly excites
- L23:2975 weakly excites
- L3:194 weakly excites
- L12:966 weakly excites
- L15:1543 weakly suppresses
- L10:1232 weakly suppresses
- L8:1347 weakly excites
- L25:2386 weakly excites
- L13:2003 weakly suppresses
- L10:1162 weakly excites
- L6:1041 weakly suppresses
- L13:1234 weakly suppresses
- L1:328 weakly excites
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
- ovsky
- ypass
- vag
- aget
- Gratis
- /pkg
- Ừ
- Gratis
- [⏎
- �
Suppresses
- esting
- bàn
- Rudd
- asty
- 练
- roperties
- Penguins
- ocos
- 一举
- 蜕