zero right after a decimal point
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. 18 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 zero that comes right after a decimal point, as in 0.01 or p < 0.01. It is less active on the digits in citation numbers, which is its weaker side. Turning it up did not change the text in a clear way.
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.
-
A single heat pulse will switch the magnetic orientation of a nanometer-sized magnetic grain in just a few picoseconds (1 picosecond is 0.000000000001 seconds). As a comparison
-
1.176-1.709) and vomiting (p < 0.001, OR = 1.969, 95%
-
Meters to millimeters (meter to mm converter) returns numbers of mm giving meters m value. Converting 1 meter to millimeter as 0.001. See also millimeters into meters calculator. See also millimeters into
-
3, 95% CI = 1.110-1.529), local citizens (p < 0.001) and officials/clerks (p = 0.00
-
) – this is one thousandth of a meter. In scientific notation, 1 millimeter is 1 – E3 m, which means 0.001 x 1
-
.91% of sampled diarrhea patients. The seasonal distribution of norovirus infections was different from non-norovirus patients (p < 0.001), with a half-year peak. Higher proportions of males (p =
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: "If the first person is a " ( " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " "
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
the way for all-optical circuits (M Mützel et al 2002 Phys. Rev. Lett. 88 083601-1).⏎The light field of an interference pattern can induce an
-
ate the molecular basis of the genetic susceptibility to alcoholism.(JAMA. 1992;268:1877-1882)<|endoftext|>The Imaginative Treatment of Geometry in Waldorf Education⏎A
-
, CT scan, rule, minor head injury⏎Ann Afr Surg. 2018;15(2):57-61 DOI:http://dx.doi.org/10.4314
-
are and where we need to go in the 21st century.⏎Longwall USA, 1998 Jun; :123-138⏎Although the number of operating longwall mining systems in the U.S.
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
- L17:259 excites
- L16:211 weakly suppresses
- L5:2309 weakly suppresses
- L17:1445 weakly excites
- L8:2526 weakly suppresses
- L9:2524 weakly excites
- L11:21 weakly excites
- L14:741 weakly suppresses
- L14:2796 weakly excites
- L17:2502 weakly excites opening parenthesis before an abbreviation
- L9:2453 weakly excites
- L16:2541 weakly excites
- L5:315 weakly suppresses
- L5:883 weakly excites
- L0:411 weakly suppresses
- L9:2401 weakly excites
- L12:2528 weakly suppresses digits inside numbers
- L7:2137 weakly suppresses
- L17:2081 weakly excites
- L13:779 weakly excites journal names after 'published in'
- L4:2386 weakly excites
- L11:1157 weakly excites
- L1:2573 weakly suppresses
- L9:2849 weakly excites
Feeds later layers
- L23:1214 weakly excites
- L19:2516 weakly suppresses
- L20:2400 weakly suppresses
- L20:1246 weakly suppresses
- L25:497 weakly suppresses
- L26:1633 weakly excites
- L23:1381 weakly excites
- L19:1653 weakly excites
- L25:2174 weakly excites
- L25:1100 weakly suppresses
- L24:2639 weakly suppresses
- L22:1341 weakly excites
- L22:449 weakly excites
- L19:2668 weakly excites
- L22:72 weakly excites
- L23:2720 weakly excites
- L26:300 weakly suppresses
- L19:1136 weakly excites
- L26:211 weakly excites
- L21:2174 weakly excites
- L22:195 weakly excites number right after a year or range word
- L19:2590 weakly suppresses
- L21:634 weakly excites
- L23:2444 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
- NDER
- рав
- strar
- stra
- знания
- ificance
- ật
- stringify
- berman
- kinson
Suppresses
- unction
- 원
- 约束
- acket
- iov
- 膑
- 卫
- 元
- 锟
- 次