comma after an introductory phrase
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. 15 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 commas that come straight after a short opening phrase, such as 'Highly visual, each topic' or 'As a community organiser, Alliah'. It is only weakly negative on the word 'but' in contrasts. Turning it up produced repetitive, looping text, and turning it down produced rambling sentences, so the steering adds little to explain it.
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.
-
and much ornamentation. A masterpiece of Juan Arellano, the bridge was completed in 1922. Paolo Alcazaren
-
the solutions to such problems. If you have an idea for an applied, data-centric note of no more than 3,000 words, please
-
-leading at a Glance series. This easy-to-read, accessible guide brings together all the key principles of caring for patients with medical and surgical needs. Highly visual, each topic is covered in one two-page spread, making it easy to quickly read
-
and Forest Ecosystems⏎Appendix 3: Estimates of Monetary Values of Ecosystem Services⏎"A landmark study on one of the most pressing problems facing society, balancing economic growth and ecological protection to achieve a sustainable future."⏎- Simon Levin,
-
see the cervix and the inside of the uterus.⏎- Laparoscopy. This minor procedure uses a laparoscope, a viewing instrument with a light, to see your uterus and pelvis. The laparoscopy is inserted through
-
.S. Fish and Wildlife Service concluded the sicklefin redhorse doesn’t need to be placed on the Endangered Species list. Though long recognized by the Cherokee, this fish was discovered by science in the early 1990s.
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, the, the, the, the, the, the, the, the, the, the, the, the, the, the, the, the,their,their,their,their,their,their,their,
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
form or of you get them in your food, that you will be healthy and go to the doctor less.⏎In countries without universal healthcare (or ones with it but with co-pays) this means that you will end up spending a lot less
-
new st<|endoftext|>Can a small camera possibly have a large lens in the front?⏎There are really "large" lenses. This search is not the biggest one, but the look of it is great.⏎But you are referring to a large diameter compared
-
9:17⏎1. Newton's model of the universe made four assumptions which were that the universe is...?⏎- Infinitely large, Infinitely old, Static and created by god⏎- Infinitely large,
-
moment of Austen’s birth or there about, John Adams left his farm in Massachusetts for the Continental Congress in Philadelphia. Doesn’t sound particularly interesting, I know but consider this.⏎John Adams left his home in mid-December 177
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
- L14:35 strongly suppresses
- L9:77 strongly excites commas in citations and name lists
- L11:154 suppresses commas after citation numbers
- L4:3 excites commas after citation fields
- L10:335 excites commas in catalog, date, and list metadata
- L5:16 excites
- L3:71 suppresses
- L0:146 suppresses
- L10:667 weakly suppresses commas inside lists of names
- L0:1419 weakly suppresses
- L13:977 weakly suppresses
- L6:798 weakly excites
- L6:620 weakly suppresses "and" in lists of three or more
- L10:144 weakly suppresses
- L2:66 weakly excites
- L1:86 weakly excites
- L5:1197 weakly suppresses
- L0:2921 weakly suppresses
- L5:5 weakly suppresses digit after a label like paper or type
- L7:2041 weakly suppresses 'or' offering an alternative
- L0:827 weakly excites
- L7:56 weakly excites
- L12:1698 weakly excites
- L6:502 weakly suppresses
Feeds later layers
- L17:118 excites comma after a clause ending
- L19:383 excites commas after a clause or list item
- L16:846 excites comma opening an aside or explanation
- L18:2863 excites
- L19:726 excites comma inside large numbers
- L23:184 excites
- L16:94 suppresses
- L18:366 excites comma inside numbers and dates
- L17:2001 excites
- L18:1394 excites
- L16:2577 excites
- L17:2677 suppresses
- L16:122 suppresses
- L17:585 excites
- L17:1841 excites
- L16:120 suppresses
- L16:17 weakly suppresses
- L16:91 weakly suppresses
- L16:558 weakly suppresses
- L16:1659 weakly excites
- L16:105 weakly suppresses
- L17:2250 weakly excites
- L16:0 weakly suppresses
- L16:66 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
- idelberg
- который
- さんは
- которое
- lette
- 特点是
- зе
- VIA
- くる
- rah
Suppresses
- ,
- 了自己的
- ”—
- 了
- inta
- exels
- elsen
- 在我
- 了些
- addCriterion