capital M inside abbreviations
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. 27 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 a capital M that starts a short name or abbreviation, such as the M in MWh, MIF or MSE, especially right after an opening bracket. When it was turned up, the model's writing turned into loops of repeated words that had nothing to do with the topic.
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.
-
s, flush residues after fossil fuels are burned and cool power plants.⏎Coal uses up to 3.2 cubic metres of water per megawatt hour (MWh) of electricity produced, gas uses up to 1.7 m3
-
lab is equipped with nuclear-powered capabilities to vaporize rocks, ingest soil, and determine whether Mars was ever inhabited by smaller forms of life. For updates follow @M
-
: livelihoods and health, education, respectable employment and financial stability, and financial security. This is accomplished through the use of the Multi-dimensional Inequality Framework (MIF). This is one of the first initiatives in Vietnam to comprehend the
-
least two minutes each time. Brush your teeth in circular strokes, and clean all surfaces, especially the backs, and chewing surfaces.⏎2. Use mouthwash:⏎Mouthwash can help destroy bacteria that cause bad breath and gum disease in your mouth
-
10 log10(MAX^2/MSE) = 10 log10(MAX/(MSE)^(1/2))^2 = 20 log10(MAX
-
resistance towards oxidation and corrosion. Besides that, volatility loss and deposit of carbon should be kept low for industrial and machinery lubrication.⏎* Malaysian Palm Oil Board (MPOB),⏎6 Persiaran Institusi, Bandar Baru Bangi
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: Likeathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathathath
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
before the Senate Agriculture Committee, which is considering the carbon-farming legislation. Introduced by senators Mike Braun (R-Indiana) and Debbie Stabenow (D-Michigan), the bill sets up the US Department of Agriculture to certify independent consultants
-
recognizing what meaning the word contributes to the passage in which it appears.”⏎For more information, The report card can be found online: www.nationsreportcard.gov.MORE IN Vermont News⏎- Most Popular⏎- Most Emailed⏎-
-
have grown, ____10___ have usually come to represent more than a single neighborhood.⏎3. Прочитайте раздел Грамматического справочника Модальные глаголы и их эквиваленты и выполн
-
of Pharmacy terminated.⏎1905-Theological Seminary organized in Waco.⏎1907-Separated from Baylor University.⏎1910-Moved to Fort Worth.⏎1918-College of Dentistry organized in
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
- L22:1138 suppresses
- L17:230 excites
- L24:1190 excites
- L10:311 suppresses
- L0:73 excites
- L21:1679 excites
- L23:270 excites
- L20:234 suppresses
- L18:553 excites
- L2:1563 excites
- L19:334 suppresses
- L7:291 excites
- L18:133 excites
- L1:16 excites
- L1:35 suppresses
- L19:623 excites
- L21:505 suppresses
- L20:2203 weakly suppresses
- L16:428 weakly suppresses unclear: mid-sentence commas, 'most' before a verb
- L26:520 weakly excites
- L4:393 weakly suppresses
- L0:844 weakly excites
- L0:1585 weakly suppresses
- L9:1040 weakly suppresses
Feeds later layers
None: this is the last layer.
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
- 干部队伍
- หมายเลข
- 要做
- 也只是
- delegated
- 不间断
- tection
- 担负
- �
- 无敌
Suppresses
- arine
- ath
- ac
- ass
- ovie
- MMMM
- anus
- AP
- ASTER
- bps