dollar sign before an amount
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. 25 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 dollar sign in front of a money amount, like '$7 million' or '$1 billion'. When it was turned up, the model's text drifted into odd, repetitive fragments, so the steering gave no clear theme. The opposite side reacts to degree signs after temperatures.
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.
-
said psychologist Jack van Honk of Utrecht University in the Netherlands, who led the study published May 24 in* Proceedings of the National Academy of Sciences*.⏎In the study, a few dozen females received half a milligram of testosterone
-
in turn protects and maintains pressure whilst keeping debris out, ensuring proper function and safety.<|endoftext|>Warner Bros./Time Warner/DC’s Wonder Woman just earned another $7 million in its seventh weekend of release, bringing the superhero movie’s domestic total
-
false positive rate as a result of mammography diagnosis results in 1.6 million unnecessary surgeries at an economical cost of over $1 billion annually in the US alone. Cyrcadia’s historic clinical results reflected
-
12 at 10:31 p.m. PDT, the world witnessed NASA’s rover landing on the Marian surface marking the next exciting phase for the $2.5 billion (USD) Mars Science Laboratory and Curiosity rover mission.
-
⏎>>> len("a string") 8 >>> >>> s = "a long string" >>> >>> len(s) 13 >>> >>> len("") 0 >>>⏎>>> >>> s3 = '' # empty string using single quotes >>> s3
-
the space agency announced that it has awarded Elon Musk’s company a contract to build a lunar lander for its Artemis mission.⏎The contract, which is worth $2.9 billion, means that a SpaceX vehicle will be responsible for putting humans
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 potential, the growth, the growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth, growth
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
In July, the warmest month, the average summit temperature is -19° C (-2° F).⏎When George Mallory, the British climber was
-
00, resulting in stepping angles of 30, 15, 5, 2.5, 2, and 1.8° per step.⏎Step angle = 360°/No of steps per
-
for crop pests. Some stinkbug species will not survive in extreme cold, and bean leaf beetles get knocked back when winter temperatures fall below 14° F for sustained periods of time. However, most soil insects are buffered by the
-
8°C (36° and 46°F). Do not freeze. Discard if the vaccine has
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
- L21:195 strongly excites
- L17:38 strongly suppresses
- L18:48 strongly suppresses
- L23:35 strongly excites
- L20:863 excites
- L21:173 excites
- L19:131 suppresses
- L14:51 excites colon that introduces a statement
- L9:7 excites
- L23:474 excites
- L16:170 suppresses
- L10:33 suppresses
- L15:123 suppresses
- L2:204 suppresses
- L14:348 suppresses
- L20:8 suppresses
- L18:267 excites
- L20:603 excites
- L21:130 suppresses
- L16:2 suppresses
- L2:761 excites
- L18:887 excites
- L19:187 excites
- L1:8 excites
Feeds later layers
- L26:365 strongly excites
- L26:56 strongly suppresses
- L27:744 strongly suppresses
- L27:1113 strongly excites
- L26:919 strongly excites
- L27:166 strongly excites
- L27:1432 strongly suppresses
- L27:462 excites
- L27:1471 excites
- L26:800 suppresses
- L26:2026 excites
- L26:3057 suppresses
- L27:1144 excites
- L26:1513 excites
- L26:174 excites
- L26:383 excites
- L27:138 excites
- L27:662 excites
- L27:660 excites
- L27:334 excites
- L27:875 excites
- L27:327 suppresses
- L26:2053 suppresses
- L27:1098 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
- #c
- #w
- [z
- @a
- {j
- [v
- [w
- {k
- {i
- [y
Suppresses
- (
- [
- -
- _
- <
- /
- *
- (?:
- 🏾
- \