opening parenthesis before an acronym
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. 24 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 open bracket '(' that comes right before a short abbreviation, like the '(ATC)' in 'acute traumatic coagulopathy (ATC)'. Its turned-down side is weak and hard to read, and the steering outputs did not clearly show this 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.
-
or in addition to, these factors [2,3]. Several terms are used in the literature to refer to this condition, including acute traumatic coagulopathy (ATC), early coagulopathy of trauma (ECT), trauma-induced co
-
disparities across and within different regions and different countries when it comes to accessing science, technology and innovation (STI) and enjoying their benefits. To address these disparities and close the existing
-
Space Exploration Technologies (SpaceX)⏎Due to the recent retirement of the space shuttle, NASA⏎now
-
died the next day. Be careful what you invent. SOURCE: Listverse<|endoftext|>The Philippine Normal University-Manila, through its Science, Technology, and Society Center (STSC) in
-
The highest ranked answers are displayed first.<|endoftext|>This language guide may help you understand some words that are commonly used when reading about sexual health, sexually transmissible infections (STIs) and blood-borne viruses (BBVs) for trans, gender
-
farin" and "Initial evaluation and management of shock in adult trauma".)⏎Coagulopathy in trauma patients, and specifically acute traumatic coagulopathy (ATC), is associated with higher transfusion requirements, longer intensive care unit and
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 first time I was in a new land, I was so afraid that I was making a mess. I was really afraid that I was making a mess, and I was afraid that I was making a mess in every possible way. I
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
same (zero) in⏎opposition to plurals (-s).⏎A small number of nouns mark plural and mass the same in
-
ous / hormonal) control is called ............... . This enables the nephrons to maintain a stable GFR even when arterial blood pressure changes.⏎Which mechanism of renal autoregulation responds to an INCREASE in BP which causes an increased STRET
-
ities very foundation of ‘being’ is rooted with the bond between man and woman.⏎Why is it so hard to get possessives and plurals right? Wait, no, that’s not my favorite part of the blog
-
, and a lack of skilled youth; all contribute to the devastation of another critical factor for a nation’s progress: a strong economy. Political instability and a paucity of skilled people de-motivate investors who are
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
- L23:290 excites
- L6:451 excites line break before a subheading
- L17:1812 suppresses
- L23:2818 excites
- L18:1747 excites
- L22:2583 weakly excites
- L21:2345 weakly suppresses
- L21:2349 weakly excites
- L18:1741 weakly excites
- L21:2364 weakly excites
- L20:2497 weakly suppresses
- L21:1761 weakly suppresses
- L23:404 weakly suppresses
- L3:2618 weakly suppresses
- L19:435 weakly suppresses
- L4:1847 weakly suppresses
- L22:1167 weakly excites
- L2:2488 weakly suppresses
- L20:995 weakly excites
- L18:1790 weakly suppresses
- L1:1900 weakly excites
- L12:1649 weakly suppresses
- L12:647 weakly excites
- L22:179 weakly excites
Feeds later layers
- L26:1004 strongly excites
- L27:1143 excites
- L27:253 weakly suppresses
- L25:2975 weakly excites
- L25:1526 weakly excites
- L25:987 weakly suppresses opening bracket before an acronym
- L27:686 weakly suppresses
- L25:281 weakly suppresses
- L27:2264 weakly excites
- L27:824 weakly excites
- L25:1986 weakly suppresses
- L25:440 weakly excites
- L26:633 weakly suppresses
- L27:792 weakly excites
- L25:2403 weakly excites
- L27:1904 weakly suppresses
- L25:2332 weakly suppresses
- L25:710 weakly suppresses
- L25:2735 weakly suppresses
- L27:369 weakly suppresses
- L27:1179 weakly excites
- L26:1705 weakly suppresses
- L27:2580 weakly suppresses
- L26:1509 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
- (pt
- Pt
- (bt
- →pt
- (ct
- ct
- -pt
- →bt
- (ht
- ct
Suppresses
- carousel
- 戡
- 稗
- TSR
- Wallpaper
- TPP
- sponge
- �
- 諮
- .yahoo