'in' before a place name
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. 21 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 the word 'in' when a country, state or city is about to follow, as in 'in Europe' or 'in Iowa'. When it was turned up, the model kept naming places like the United States over and over.
Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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.
-
voltage drop waveforms are in-phase with each other.⏎Given that the power supply frequency is 60 Hz (household power frequency in the United States), calculate impedances for all components and determine all voltage drops using
-
food processing plants, accounting for 88 million tons of municipal waste each year in Europe. The Commission notes that the main environmental threat from bio-waste is the
-
now have customized climate predictions, thanks to a<|endoftext|>What is redistricting and how does it work in Iowa? Your questions answered.⏎The U.S. Census Bureau delivered its long-delay
-
and playing duck-apple at Halloween; burning the clavie (tar barrel) at new year or 'Hallooing Largesse' (where, in
-
, usually within three hours. The results varied depending on proximity to the natural mating season. For L. guildingii, the peak was in mid-summer in Australia (December). (Charonia Research, 1996b;
-
rather as a non-British symbol to counter the English St. George societies. This St. Nicholas society was similar to the Sons of St. Tammany in Philadelphia. Not exactly St. Nicholas
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 number of $ (3, 3, 2) $ on the $ (3, 3, 2) $-categorical $ 3 $-dimensional $ (3, 3, 2) $
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
also called Zoanthidea or Zoanthiniaria) are an order of cnidarians commonly found in coral reefs, the deep sea and many other marine environments around the world. These animals come in a variety of different colonizing formations and in
-
So fill up that page and let your imagination run WILD!<|endoftext|>Baleen and sperm whales, known collectively as the great whales, include the largest animals in the history of life on Earth. With high metabolic demands and large populations, whales
-
to consider where art comes from and how it impacts all of us.⏎Contributors to this open textbook offer relatable descriptions, details, and contexts of artworks from around the
-
may be moderated or removed.<|endoftext|>Even the hardy tardigrade might not stand up to climate change⏎Tardigrades are well-known for being one of the hardiest animals on the planet, but even they have their limits. New
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
- L18:2225 strongly suppresses
- L18:1099 strongly excites
- L19:2957 strongly excites
- L19:160 strongly suppresses
- L19:2170 suppresses
- L18:1861 excites
- L18:2899 excites
- L19:2778 suppresses
- L18:2980 suppresses
- L19:2859 excites
- L17:1178 excites
- L20:1288 excites
- L17:596 suppresses
- L18:1860 suppresses
- L14:1550 suppresses
- L16:1968 excites
- L18:183 suppresses
- L15:1464 suppresses
- L17:2267 suppresses
- L16:2771 excites
- L16:783 suppresses
- L17:1167 excites
- L20:691 suppresses
- L17:897 suppresses
Feeds later layers
- L27:536 strongly excites
- L22:1843 excites
- L27:1201 excites
- L22:2433 excites
- L23:1087 excites
- L26:207 suppresses
- L23:2950 excites
- L25:2455 excites
- L26:825 excites
- L26:559 excites
- L24:893 excites
- L22:439 suppresses
- L27:405 excites
- L27:2537 excites
- L24:1851 excites
- L25:3071 excites
- L23:1248 excites
- L22:2246 excites
- L26:1751 excites
- L24:2306 excites
- L27:1130 excites
- L24:993 excites
- L22:2439 excites
- L26:2446 excites
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
- 中国的
- 在中国
- 在美国
- 中国大陆
- 在日本
- 在我国
- 中国
- 美国
- France
- China
Suppresses
- ming
- ún
- 冊
- Plot
- 绰
- _mar
- 날
- ulta
- wer
- unik