'to' before a verb of purpose
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. 22 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 lights up on the word 'to' when it comes before an action, as in 'how to calculate' or 'used to find'. It is predicting that a verb like 'find' or 'calculate' comes next. When it was turned up, the model got stuck repeating 'find' and 'a good way to find'.
Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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.
-
should be called getVolumeOfSphere(), take one parameter called radius and apply the relevant formula to calculate and return the volume of a sphere based on the given radius.⏎- Your
-
(9/10)k-2, where k is the number of digits of 2n. The number of digits can be estimated as n log102. Thus, the probability is about cxn
-
do the rest, including converting the decimal to a percentage.⏎To find the percentage value of a mark, you’ll need to know its total mark. In order to find this, use basic arithmetic. Divide the mark by the highest possible mark and
-
.net, a new website, provides goal writing guidance, especially useful for early intervention and children on the autistic spectrum disorder.<|endoftext|>The online WACC Calculator is used to calculate the weighted average cost of capital (WACC).⏎In finance, The weighted
-
a deep breath and let us take it one step at a time. Click and begin.⏎GCF and LCM by Prime Factorization⏎A straight forward way to find the greatest common factor (GCF) and least common multiple (LCM
-
important, if not the most important, metrics in valuing the bond.⏎Now, let's look at some examples to understand how to calculate bond prices.⏎How to calculate the bond price? The bond price formula⏎Bond price is calculated as 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: I have a 2000 model, 2000 2000, 2000, 2000, 2000, 2000, 20
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
and the slope. The discrete model of exponential growth is ; where the first term is , and is the common ratio. The continuous model is .⏎1. Find the common ratio of each geometric sequence.⏎
-
5 Y 2 . Solution : X =- 5 W c) Find some solution of A X = 3 Y 1- 5 Y
-
It is true that today we have online limit calculators that can help in limit calculation. In this post we are going to discuss the reasons as to why find the limit calculation is a tough task especially for high school students.⏎Teaching math to
-
x x = 2. 1 cos lim x x x -= 3. ( ) 1 lim 1 x x x e + = Strategy for Finding Limits: 1. Try direct substitution: One of three things will happen:
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
- L21:234 suppresses
- L21:2199 suppresses
- L16:2597 excites
- L4:3053 suppresses
- L20:700 excites
- L18:1579 excites
- L17:2903 excites
- L21:1651 weakly excites
- L21:317 weakly excites
- L1:710 weakly suppresses
- L18:2999 weakly excites
- L17:2257 weakly suppresses
- L1:2206 weakly suppresses
- L5:78 weakly excites
- L17:789 weakly excites
- L15:1445 weakly excites
- L15:2437 weakly suppresses opening parenthesis after a noun
- L17:1912 weakly excites
- L18:2584 weakly suppresses
- L18:2628 weakly suppresses
- L4:713 weakly excites
- L12:1390 weakly excites
- L12:354 weakly excites
- L13:2110 weakly excites
Feeds later layers
- L27:1284 strongly excites
- L24:1909 excites
- L23:540 excites
- L27:1207 excites
- L26:2988 excites
- L25:660 excites
- L24:1302 excites
- L24:769 excites
- L27:2510 excites
- L23:1325 excites
- L23:1432 excites
- L23:494 excites
- L26:1422 weakly excites
- L24:837 weakly excites
- L25:1335 weakly excites
- L24:1068 weakly suppresses
- L25:2811 weakly excites
- L24:2247 weakly excites
- L26:1282 weakly excites
- L26:656 weakly suppresses
- L23:3066 weakly suppresses
- L24:545 weakly excites
- L23:136 weakly excites
- L25:1035 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
- finding
- 计算
- find
- find
- computing
- Find
- finding
- 求
- finds
- Find
Suppresses
- pered
- AFX
- brush
- 听取
- izzly
- 回头看
- 诃
- abyte
- serialVersionUID
- مل