'to' starting an infinitive
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. 9 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 fires on the word 'to' in phrases like 'threatens to', 'tend to', or 'is to be', where a verb is about to follow. The opposite side fires on 'to' in phrases like 'want to know' or 'glad to see', so the split is not clean. Turning it up made the text drift and repeat.
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.
-
applications and it makes me sad. Little curiosities like this can grow into amazing large-scale processes with a bit of support, but a lack of funding threatens to take that away. A public conditioned to a perpetual hype/fatigue cycle are
-
aircraft from flying into the sort of engine-seizing ash cast up by Eyjafjallajökull, the low-lying volcano whose plume all but shut down
-
we (people) can know objectively and comprehend the objects around us. (146)⏎forms of understanding from which reality is to be deduced. (146)⏎refers to the philosophical movement and
-
they have with their family, community, culture and place.⏎Being is about living here and now. Childhood is a special time in life and children need time to just ‘be’- time to play, try new things and have fun.⏎B
-
movements needed to do this are controlled by special neuron bundles in the spinal cord. It is not quite clear how these central pattern generators know how quickly the legs are to be moved.⏎An international team working with scientists from Harvard University and the Max Plan
-
during illness or with a busy schedule.⏎- ‘Cooking’ up excuses to not eat.⏎- Wears baggy clothes to hide the body.⏎- Going to the bathroom right after meals to throw up the food just eaten.⏎- People with
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 like to have been thinking about think about do something get to go go go go go go go go go go go go go go go go go go go go go go go go go go go go go go go go go go go
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
also keep them from telling us how they feel.⏎On the other hand, it also isnt wise to confront children with information that they may not yet understand or want to know. As with any sensitive subject, we must seek a delicate balance that encourages
-
of that era helped push the reforms that redefined how we live and how much misery we will accept.⏎It was a bargain that most of us would be glad to see return. The problem with today
-
were gone."⏎Join to answer this question⏎Join a community of thousands of dedicated teachers and students.Join eNotes<|endoftext|>Is there something else you would like to know?⏎Use the alphabetical glossary listing to find a short definition to a medical
-
the state of⏎Andhra Pradesh, it is more comfortable to wear it over a shirt than as a single⏎piece of clothing.⏎Q4. What do you wear on veshti?⏎<|endoftext|>Copyright (c) Arvin S. Quist
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
- L6:1880 strongly excites "by" introducing a deadline or means
- L6:914 strongly excites
- L6:439 suppresses
- L6:1133 excites "to" before an infinitive verb
- L8:369 suppresses 'of' after 'learned' or 'because'
- L6:2747 suppresses
- L7:681 excites 'from' after prevent or stop
- L4:113 suppresses
- L7:751 excites
- L6:2359 excites
- L8:403 excites
- L8:94 excites
- L6:751 suppresses
- L6:2057 suppresses
- L6:2020 excites
- L6:2064 suppresses
- L4:518 suppresses
- L6:1621 suppresses
- L1:906 suppresses
- L8:1288 suppresses
- L8:723 suppresses
- L7:1505 suppresses 'the' after a verb or preposition
- L7:14 excites
- L4:2890 suppresses
Feeds later layers
- L16:1153 strongly suppresses
- L14:659 strongly suppresses
- L17:1592 strongly suppresses
- L11:429 strongly suppresses
- L19:385 excites
- L19:776 suppresses
- L18:452 suppresses
- L20:480 suppresses
- L17:187 suppresses
- L24:208 suppresses
- L18:141 excites
- L19:617 suppresses
- L19:876 suppresses
- L19:652 suppresses
- L18:715 suppresses
- L17:1994 suppresses
- L16:1291 suppresses
- L15:1234 suppresses
- L26:1122 suppresses
- L18:472 suppresses
- L12:434 suppresses
- L15:1176 suppresses
- L19:1830 suppresses
- L20:241 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
- precedence
- bad
- superior
- morale
- enuity
- curacy
- sensible
- idence
- 最佳
- ことに
Suppresses
- /she
- astics
- 把她
- ски
- 把你
- rics
- unravel
- (Android
- aska
- brities