technical terms mid-phrase, mostly engineering
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 mixed: they peak on different tokens. 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 pieces of technical terms in engineering, chemistry and biology writing, such as 'Mill' in 'End Milling', 'Weld' in 'Welding', or 'me' in a gene modification code. It goes slightly negative on the middle of words like 'servos' and 'with' after 'interact'. Turning it up or down did not give a clear, shared output theme.
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.
-
Angle in End Milling⏎End mills are used to form shapes and holes during milling and can cut
-
, processes that regulate gene expression. They found about 80 percent of pediatric hindbrain ependymomas had substantially reduced levels of H3K27me3, a critical histone H3 protein modification. These tumor samples consistently tied
-
units, type ‘UN’ into your keyboard and then select the ‘ENTER’ key. A dialog box should pop-up on the screen allowing you to specify the unit
-
point (2, 4)?⏎November 18, 2010⏎biology / plasmolysis of human red blood cells⏎Plasmolysis of human red blood cells would occur if the cell were in .. solution A
-
basic knowledge about rake angle and introduces the latest technology used to take accurate rake angle measurements.⏎- Rake Angle⏎- Positive and Negative Rake Angle in End Milling⏎- Relationship Between Cutting Tool Angle, Rake Angle, and Relief Angle
-
transfer commonly used on low alloy steels. This all position wire normally used in appliions which will produce a⏎Penetration Form of DC-TIG Arc Welding Using Flux
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 the color blue, the color green, and the color yellow." What does that mean? Answer: [Insert image] The answer is to use the image or a link. Answer: [Insert image] The answer is
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
a wide variety of things - it doesn't have an Operating System (OS) per se. Arduino has a great deal of support for accessing the board and driving servos, etc., but it is limited and doesn't provide what an OS provides
-
ASPBERRY Pi and ARDUINO COMPUTERS⏎The Arduino is a wonderful platform for handling the lower level work of a robot - reading sensors, driving servos, and blinking lights. There is a fair amount of RAM on board, and
-
is a solution.⏎The program svn, which is used on the command line interact with the SVN server, has a diff function too: svn diff.⏎You can run
-
'S', then click at the end of the word and type<|endoftext|>Mallorca’s popular tourist railway
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
- L4:2279 excites
- L6:989 excites
- L8:676 suppresses
- L8:12 excites
- L8:1005 suppresses
- L7:2730 excites
- L6:18 excites
- L6:1206 weakly suppresses
- L7:2764 weakly excites
- L8:7 weakly excites
- L7:1980 weakly suppresses
- L8:4 weakly excites
- L0:2376 weakly suppresses
- L6:1795 weakly suppresses
- L6:438 weakly suppresses
- L8:1542 weakly excites
- L7:2915 weakly excites
- L7:1212 weakly excites
- L6:827 weakly suppresses
- L0:1057 weakly suppresses
- L6:2881 weakly suppresses
- L8:13 weakly suppresses
- L3:439 weakly excites
- L6:7 weakly suppresses
Feeds later layers
- L10:1675 suppresses
- L21:2025 suppresses
- L10:2446 suppresses
- L10:2806 excites
- L13:2456 excites
- L10:2777 suppresses
- L21:3021 suppresses
- L10:1629 excites
- L10:2544 excites
- L11:96 weakly excites
- L17:1534 weakly suppresses
- L11:976 weakly excites
- L17:2028 weakly excites
- L11:1029 weakly suppresses
- L20:953 weakly suppresses
- L12:1223 weakly excites
- L14:2775 weakly suppresses
- L14:2447 weakly excites
- L18:2943 weakly suppresses
- L16:2589 weakly excites
- L11:2037 weakly excites
- L20:1628 weakly excites
- L10:1996 weakly suppresses
- L25:1764 weakly 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
- arium
- 了自己的
- elian
- 储备
- childhood
- 保税
- lesia
- amphetamine
- październ
- spreading
Suppresses
- tempts
- rome
- testers
- カメ
- беж
- locate
- APTER
- ATER
- tester
- ients