the year '2030' ending in zero
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. 3 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 final digit '0' in the year 2030 or similar years like 2040 or 2050, when the year is written as a target date in the future. Turning it up produced loose, unrelated text.
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.
-
is very familiar with the role of standards. Well-known standards, such as ISO 17799 for IT security and the ISO 9000 series for management, establish guidelines and general principles that codify industry best practices.
-
of the oceans is widely regarded as a possible solution to increase food and could be one of the next most important developments in human history. By 2050, the edible bioresource biomass will have to satisfy the 9 billion people predicted
-
production will be the most affected by climate change by 2030. This would lead to food insecurity, which in turn could lead to social instability
-
its commitment to meeting Sustainable Development Goal (SDG) 6, which is to “ensure access to water and sanitation for all” by 2030. But the National Longitudinal Phone Survey (NLPS) on the socio
-
a plan that the US has to send humans to Mars by 2030.” Space is about to get crowded, which means it’s even more important not
-
jobs and investments in frontline communities bearing the brunt of environmental injustice—they offer a blueprint for both cutting U.S. carbon emissions in half by 2030 and doing so equitably.⏎But neither plan explicitly grapples with the politically
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: what is the value of the number of the number of the number of the number of the number of the number of the number of the number of the number of the number of the number of the number of the number of the number of the
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
versal of dysbiosis as a therapeutic model in MS⏎- Roles for short chain fatty acids, bile acids, phytoestrogens, tryptophan, choline and mucin degradation⏎And for those readers interested in data
-
almonds instead of unhealthy foods and lose weight by eating the tasty treat.⏎Another of the health benefits of almonds are good for brain development. The content of tryptoph
-
86 from a portion of Ancient Falmouth. The latter was later to be dissolved into Cape Elizabeth, Westbrook, and an additional portion of Portland, Stroudwater.⏎On March 26, 1833 Portland became a
-
sensitivity to drought. Genotypes will be subjected to drought by adding PEG 8000 to the nutritive solution (osmotic potential of -0.75 MPa). In addition to evaluating the differential production of NO by
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
- L2:148 weakly excites
- L2:1019 weakly suppresses
- L0:207 weakly suppresses
- L2:2646 weakly suppresses
- L2:1246 weakly suppresses
- L2:126 weakly excites
- L2:111 weakly suppresses
- L1:2507 weakly excites
- L2:41 weakly suppresses
- L0:2473 weakly excites
- L0:1301 weakly suppresses
- L0:1111 weakly excites
- L1:2200 weakly suppresses
- L2:1947 weakly excites
- L0:178 weakly suppresses
- L0:1762 weakly suppresses
- L2:879 weakly suppresses
- L1:1921 weakly suppresses
- L1:1734 weakly excites
- L2:328 weakly excites 'of' in fixed names and phrases
- L0:1224 weakly suppresses
- L1:2263 weakly suppresses
- L1:2010 weakly excites
- L1:1757 weakly suppresses
Feeds later layers
- L7:2273 weakly suppresses
- L16:2492 weakly suppresses
- L6:1398 weakly excites
- L5:542 weakly excites
- L22:17 weakly excites
- L4:468 weakly suppresses "that" after "than" or "like"
- L14:1527 weakly suppresses
- L4:2051 weakly suppresses
- L17:2958 weakly excites
- L18:729 weakly suppresses
- L17:606 weakly excites
- L16:1896 weakly suppresses
- L21:1378 weakly excites
- L14:1547 weakly suppresses
- L11:2180 weakly suppresses
- L18:1879 weakly suppresses
- L23:2604 weakly suppresses
- L10:2840 weakly suppresses
- L14:2226 weakly suppresses
- L8:2796 weakly suppresses
- L18:2931 weakly suppresses
- L6:2435 weakly excites
- L16:644 weakly excites
- L6:289 weakly 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
- 阴影
- 床
- inished
- culos
- оцен
- gratuit
- Boom
- 这种方法
- 'é
- pis
Suppresses
- SID
- oram
- ['./
- onDataChange
- elpers
- су
- DISCLAIMED
- []):
- []);⏎
- 又好