blank slot before a founding year
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. 8 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 spot right after 'founded in' or 'was founded in', where a year is about to come next. Turning it up pushed the model toward unrelated text about math and weather, which does not show the same idea.
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.
-
Thegsun Tashi Gomang⏎The Karma Thegsun Tashi Gomang ws founded in 1980. The Gyalwa Karmapa (a Buddhist priest
-
Ballet of the Twin Cities; Immigration and Refugee Services of America; and the International Institute. Founded in 1919, the International Institute was established to provide various services for migrants
-
museum has a scientific staff of more than 200, and sponsors over 100 special field expeditions each year.⏎The Museum was founded in 1869 and housed in the old Arsenal building in Central Park. Theodore
-
the 'Hereford Herd Book Society', founded in 1878.⏎Polled Herefords were developed from the horned
-
around 1000. The monastery belonged, as does the current church, to the Benedictine Arch Abbey of Pannonhalma which was founded in 996. Around 1550 the monastery was
-
. The Society is affiliated to the European Public Health Association and is a member of the International Epidemiological Association’s European Epidemiology Federation.⏎The Society was founded in 1956.<|endoftext|>- THE MAGAZINE⏎- NEW PRODUCTS⏎
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 fact that the present experience is a part of the current state of the previous thought. So, the present experience is a way to get to the present, and the present experience is a way to get to the present. So, the
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
use of the phrases hope, love, dwelling, college, youngsters, group, help, and assist after the drills. The researchers thought of posting with these phrases within the aftermath of the drills to be indicative of getting excessive nervousness.⏎They included
-
will be used to approximate all real zeros of a polynomial.⏎This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.⏎Spreadsheets: Laying a Foundation for Understanding Functions,⏎Spreads
-
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document⏎Unformatted text preview: itself, and if ( ) ( ) lim lim x
-
red flag.⏎Copyright (c) 2022 Proceedings Of International Conference On Communication Science⏎This work is licensed under a Creative Commons Attribution 4.0 International License.<|endoftext|>How do you explain how babies are made to an 8
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
- L6:72 suppresses
- L5:191 excites
- L6:1893 weakly excites
- L7:2752 weakly excites
- L7:1272 weakly suppresses
- L5:1363 weakly excites
- L7:2173 weakly suppresses
- L7:383 weakly excites
- L7:707 weakly excites
- L5:290 weakly excites
- L4:74 weakly suppresses
- L1:2394 weakly suppresses
- L1:118 weakly suppresses
- L0:984 weakly suppresses
- L4:861 weakly excites
- L4:1419 weakly suppresses
- L6:659 weakly excites
- L4:428 weakly excites
- L6:194 weakly excites
- L5:1455 weakly excites
- L4:2136 weakly excites
- L4:810 weakly suppresses
- L5:940 weakly suppresses
- L4:2292 weakly excites
Feeds later layers
- L10:2021 suppresses
- L14:523 suppresses
- L10:407 excites
- L14:1230 suppresses
- L13:929 excites
- L12:1791 suppresses
- L16:2652 suppresses
- L14:2360 weakly excites
- L14:3017 weakly suppresses
- L19:307 weakly suppresses
- L13:2528 weakly suppresses
- L13:1285 weakly suppresses
- L9:1759 weakly suppresses
- L9:426 weakly suppresses
- L16:2450 weakly suppresses
- L19:1431 weakly excites
- L10:1882 weakly suppresses
- L13:987 weakly suppresses
- L13:2262 weakly excites
- L15:814 weakly excites
- L10:1320 weakly suppresses
- L21:425 weakly suppresses
- L21:3005 weakly suppresses
- L17:1093 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
- 这家伙
- dataList
- itemList
- Analysis
- Analysis
- odelist
- gente
- ño
- 硌
- подоб
Suppresses
- lete
- ikt
- ermann
- reck
- amma
- kap
- лот
- hosts
- quisite
- gae