'to' after a noun, before a purpose verb
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. 15 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' when it comes after a noun and tells you what that noun is for, as in 'something to latch on to', 'nothing to talk about' or 'a place to rest'. Turning it up made the model repeat 'to watch' and 'to feel' over and over.
Written by Claude from this neuron's activations and steering outputs. Confidence: high. Think it is wrong? Flag it (opens an email to the maker with the neuron already filled in).
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.
-
after the larvae are spawned, they reach a period of metamorphosis and go to the bottom of the tank to attach themselves to shell fragments so they have something to latch on to as they begin to grow their own shells. The farmer said any
-
atypical from the others because of that. In a way if you don’t share your memories, people will find you boring and dull, because you have nothing to talk about, because everything you talk about is a memory. A person cannot talk
-
is full of water but capped off on both ends.⏎You hit the power switch on the pumps, what happens?⏎The answer is nothing, there is no where to move the water to, the pumps don't even spin. (ignore water turbulence
-
in the well-being of the surrounding environment. Pollinators are in trouble due to habitat loss, pesticide use and the increasing threat of climate change. With nowhere to build their homes and no plants to pollinate, honeybee
-
school is a place to learn and that learning is COOL! Constantly point out things they have learned and things they have yet to learn to give them something to look forward to. Feel free to talk about your own learning experiences with them.⏎
-
the hospital⏎To provide these carers with a place to rest while staying at the hospital to look after their sick relatives, the Foundation helped
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 fact that, if there are not enough resources (i) the time and financial constraints (i) the time and financial constraints, and (i) the time and financial constraints) (i) the time and financial constraints) and (
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
well as calculating summary statistics such as means, frequency, and percentiles. Creating filters and joins on in-memory data are also discussed.Learn how to⏎Who should attendExperienced programmers who need to learn the syntax and functionality of the SAS LAS
-
Dosing InformationThe dose of Monopril your healthcare provider recommends will vary depending on a number of factors, including:⏎- Your age⏎- The medical condition being treated⏎- Other medical conditions you may have⏎- Other medications you may currently
-
the decimal and cents to the right of the decimal. To add decimals, simply line up the decimal points of the amounts of money that are being added. When you have added the amounts of money, be sure to place the
-
help maintain the nation's natural gas transmission infrastructure through the timely and effective detection of natural gas leaks through evaluation of geobotanical stress signatures.⏎The remote sensing techniques being developed employ advanced spectrometer systems that produce visible and near infrared
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
- L9:2312 excites 'to' after a noun (a role to play)
- L7:1974 excites
- L12:2519 suppresses
- L6:2297 excites
- L7:2227 excites
- L14:2265 excites
- L11:1811 suppresses
- L14:2416 suppresses
- L13:1686 suppresses
- L14:1854 excites
- L4:470 suppresses
- L6:958 suppresses
- L5:1030 excites
- L14:2791 excites
- L13:134 excites
- L13:2057 suppresses
- L12:1542 weakly suppresses
- L14:492 weakly suppresses
- L6:2182 weakly excites
- L12:1084 weakly suppresses
- L5:439 weakly suppresses
- L7:2553 weakly suppresses
- L1:464 weakly excites
- L5:1428 weakly suppresses
Feeds later layers
- L16:2537 excites
- L17:2067 excites
- L16:2555 excites
- L19:1235 excites
- L19:2757 excites
- L18:2703 excites
- L18:2440 excites
- L19:2010 excites
- L21:2805 excites 'the' right before 'subject' or 'source'
- L18:2666 excites
- L23:1542 excites
- L16:1700 excites
- L18:1672 excites
- L17:2575 excites
- L19:2621 excites
- L18:2740 excites
- L21:2602 excites
- L22:2402 excites
- L17:1884 excites
- L22:1422 excites
- L17:2756 excites
- L16:2474 excites
- L22:429 excites
- L16:2214 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
- 可供
- to
- holds
- 可
- which
- hold
- ried
- otted
- to
- which
Suppresses
- alley
- 家伙
- ouce
- SharePoint
- iences
- 噜
- uem
- SystemService
- 缯
- <num