'if' in conditional asides
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. 11 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 word 'if' when it adds a condition at the end of a phrase, like 'Discard if the vaccine has been frozen' or 'printed out if required'. Turning it up gave text that was not clearly about conditions, so the steering results do not back up the pattern.
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.
-
application of color or tone.⏎14. Gen. Arrangement Drawing⏎Exterior view all components using line only and with sufficient detail to produce an Appearance Model if required. Usually drawn in third angle projection.⏎15. Detail Drawing⏎Contains
-
fields), not just specialists like Gonthier. His talk tomorrow, titled “Univalent Foundations,” will describe this, if he makes it in time<|endoftext|>New research leads to call for ban on hands-free
-
8°C (36° and 46°F). Do not freeze. Discard if the vaccine has been frozen.⏎A separate sterile syringe and needle must be used
-
, you can leave feedback. This video shows how to do this in Brightspace:⏎As you can see, a pdf serves double duty. Hand out hard copies if you wish… or upload them digitally so students can record their responses and submit them
-
of the whole of the original book. This CD has been bookmarked for easy⏎navigation, and pages can be searched, browsed, enlarged and printed out if required.<|endoftext|>IN-clusively MindedWe have signed our names to the
-
On what type of boundary was this?⏎Where was the most recent earthquake closest to you? Include the name of the city and state you reside in (or country if outside of the United States). Was it associated with a boundary?⏎Was there any
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 problem of the United States and the United Nations. The problem of the United Nations is a series of international negotiations that will involve a global effort in the design of a set of rules, which will be used to produce a set of global
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
will walk in all the way which I command you, that it may be well with you.'⏎Matthew 7:24-27 “Therefore everyone who hears these words of Mine and acts on them, may
-
have any photos of this monument yet. Why don't you be the first to send us one?⏎If Google Street View is available, the image is from the best available vantage point looking, if possible,
-
nests, but the Son of Man has no place to rest."⏎American Standard Version⏎And Jesus said unto him, The foxes have holes, and the birds of the heaven have nests; but the Son of man hath not where to lay his
-
Holman Christian Standard BibleJesus told him, “Foxes have dens, and birds of the sky have nests, but the Son of Man has no place to lay His
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
- L9:306 suppresses
- L6:2134 excites
- L4:265 suppresses
- L6:1881 excites
- L3:749 excites
- L9:2587 excites
- L4:1093 excites
- L0:1470 excites
- L10:358 weakly suppresses
- L9:1358 weakly excites
- L6:1925 weakly excites
- L0:1920 weakly excites
- L8:1866 weakly suppresses
- L8:691 weakly excites
- L9:3051 weakly suppresses
- L10:310 weakly excites
- L6:1203 weakly suppresses
- L5:1447 weakly suppresses
- L5:84 weakly suppresses
- L10:1241 weakly suppresses
- L4:2494 weakly suppresses
- L4:1834 weakly suppresses unclear: time spans and 'once you've'
- L8:2108 weakly excites
- L2:2765 weakly suppresses
Feeds later layers
- L15:513 excites
- L16:2167 excites
- L17:619 excites
- L16:2929 excites
- L15:2342 excites
- L16:2550 excites
- L17:1195 excites
- L19:2348 excites
- L13:696 excites
- L20:2066 excites
- L18:2311 excites
- L18:1606 excites
- L18:2235 excites
- L18:2778 excites
- L14:395 weakly excites
- L23:2705 weakly suppresses
- L18:2263 weakly excites
- L16:2323 weakly excites
- L16:2838 weakly excites
- L21:919 weakly excites
- L18:2783 weakly excites
- L16:1342 weakly excites
- L17:2729 weakly excites
- L19:2573 weakly excites "of" after a percentage or share
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
- emode
- 给您
- Globals
- そうで
- 您
- oxid
- 不利
- stein
- 你不
- _URI
Suppresses
- rics
- theirs
- razón
- atas
- görünt
- 大地
- ly
- 麼
- tempt
- LY