abbreviation periods after initials
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. 0 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 full stop or comma that comes right after a short abbreviation or a single initial, like 'M.' in a name or 'U.S.' or 'D.C.'. When it is turned up, the model's writing drifted into rambling and odd phrasing, with no clear 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.
-
that threatens the sustainability and resilience of our fishing communities,<|endoftext|>Suicide and Social Change in China⏎- Cite this article as:⏎- Phillips, M.R., Liu, H. & Zhang, Y. Cult Med Psychiatry (199
-
., Swarup R., Tassinari A., Pandey B.K., York L.M., Atkinson B.S., Sturrock C.J., Mooney S.J., Hochholdinger F.,
-
the Western states? Does Nevada have a surplus?⏎Mike Cox, big game biologist, Nevada Department of Wildlife<|endoftext|>|Written by health professionals: Doctor, M.D., CphT, and others - but easy to read.|⏎METABOL
-
Bovina R., Pattem J.K., Borkar A.N., Lombardi M., Forestan C., Milner S.G., Davis J.L., Lale A., Kirschner G.K
-
.D., assistant professor of psychology in the School of Science at Indiana University-Purdue University
-
influence of aerosols on the global climate.⏎Ault, A.P., et al. “Raman Microspectroscopy and Vibrational Sum Frequency Generation
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 have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
or ‘auspicious’ (cf. Nandi verses in Sanskrit drama). Nandi is the name of Lord S
-
ada (c. 1475 or 1486 – between 15
-
period. We used repeated-state (i.e. abundance trends) and rate (i.e. colonization and persistence) responses to advance
-
cf. canguzua [In] Sea Slug Forum. Australian Museum, Sydney.
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
None: this is the first layer.
Feeds later layers
- L3:1224 excites
- L1:1925 suppresses
- L5:896 excites
- L2:1830 excites
- L3:1784 weakly excites
- L8:854 weakly excites
- L9:332 weakly excites
- L3:2089 weakly excites
- L21:1255 weakly suppresses
- L15:288 weakly excites
- L26:2577 weakly suppresses
- L5:456 weakly suppresses
- L2:2214 weakly excites
- L8:1131 weakly excites "your" addressing the reader
- L12:1093 weakly excites
- L9:1176 weakly suppresses
- L2:338 weakly excites
- L18:2478 weakly excites
- L8:1219 weakly suppresses
- L19:2792 weakly excites
- L6:58 weakly excites
- L24:102 weakly excites
- L10:2317 weakly excites
- L4:662 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
- *
- $\
- bles
- '?
- ус
- "?
- >
- acon
- -validator
- лей
Suppresses
- riday
- Perc
- getEmail
- getCell
- 一手
- ordova
- imeline
- iskey
- 客气
- umblr