the word 'without'
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. 24 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 "without" when it comes after a clause, as in "without fail", "without doubt" or "without warning". Turning it down or up did not change the generated text in any obvious way.
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.
-
intent should never be considered, ever. To even look down that road is foolish.⏎But consider this: if you walk into any contemporary art museum you will, without fail, find on the wall next to the art, a statement from the artist
-
.<|endoftext|>Though buckwheat is certainly one of the lesser known healthy grains out there, we’re sure that you’ll have had it at least once or twice without even
-
dance has had the ability to transcend all borders of language, geography and age. It is a primal urge, a surge of energy, an uplifting feeling and, without doubt, a gift.<|endoftext|>June 8, 2000⏎
-
<|endoftext|>The eyes are the windows to the soul, or so goes the proverb. As humans, we use our eyes to express a wide variety of emotions, often without even realizing we’re doing
-
this method will help you achieve a bouncing ball animation in probably 5 minutes or less.⏎So without wasting much time, let’s get into it!⏎Here are the steps:⏎-
-
at the head of the bay behind several small islands. It was a nice sheltered spot. It would even be safe from the occasional sudden great storms that struck without warning in that late summer season. As the men heaved in the anchor and
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: If the fact that certain parts of the world at certain time periods about at certain places at certain things at certain things about certain things at certain time periods about at certain time periods about at certain time periods about at certain time periods about at certain
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
enteritis. A bacteria called clostridia was isolated as the organism responsible.⏎Necrotizing enteritis is associated with poor hygiene and an inadequate diet, and it was found that
-
γ-proteobacterium, W. melaphagi is a rhizobacterium).⏎Being present endosymbiotically in 20+% of
-
Library Board Images Online⏎Anthony Farrington Trading Places, The East India Company and Asia 1600-
-
|• Land||576 sq mi (1,490 km2)|⏎|• Water||83 sq mi (210 km2) 13%%|⏎|• Density||401/sq mi
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
- L22:390 excites
- L20:692 suppresses
- L20:1414 weakly excites
- L8:2442 weakly suppresses
- L4:1711 weakly suppresses
- L19:1068 weakly excites
- L19:1513 weakly suppresses
- L9:950 weakly suppresses
- L7:2404 weakly excites
- L23:206 weakly suppresses
- L18:2133 weakly suppresses
- L23:493 weakly excites
- L17:1740 weakly excites
- L14:1468 weakly suppresses
- L1:2109 weakly suppresses
- L17:443 weakly excites
- L15:2572 weakly excites
- L0:1338 weakly suppresses
- L20:335 weakly excites
- L4:2909 weakly suppresses
- L7:2396 weakly excites
- L14:3070 weakly suppresses
- L16:752 weakly excites
- L21:1060 weakly suppresses
Feeds later layers
- L25:2783 suppresses
- L26:899 weakly excites
- L25:671 weakly excites
- L25:2953 weakly suppresses
- L25:510 weakly excites
- L25:2575 weakly suppresses
- L25:2980 weakly excites
- L27:434 weakly suppresses
- L26:2686 weakly excites
- L25:392 weakly suppresses
- L25:1645 weakly excites
- L26:1070 weakly excites
- L25:1583 weakly excites
- L25:1635 weakly suppresses
- L25:1491 weakly excites
- L25:605 weakly excites
- L27:2165 weakly excites
- L25:2505 weakly suppresses
- L26:2570 weakly excites
- L26:2449 weakly excites
- L25:484 weakly excites
- L26:754 weakly suppresses
- L27:2604 weakly excites
- L27:1533 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
- ))))
- ])))
- )').
- }})⏎
- ))).
- }`).
- )").
- "))))⏎
- ))));⏎
- )))),
Suppresses
- ắn
- isky
- �인
- LEGAL
- ника
- gings
- orage
- ности
- ник
- 痣