'in' in 'keep in mind'
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 fires on the word 'in' in phrases like 'bear in mind' and 'keep in mind', where 'in' comes right before 'mind'. Turning it down made the model repeat the word 'in' over and over, which shows how strongly it is tied to this one small word.
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.
-
them fascinating. After you have mastered some fundamental activities, you’ll be able to try new ones and incorporate a brand new twist for the previous kinds. And bear in mind, it’s in no way too early to start creating new memories. >>>
-
forest for one use if it's going to compromise other uses? Our forests are incredibly important for harvesting but also for tourism [and] we want to always bear in mind the possible tradeoffs.⏎Tourism is a huge asset to Maine
-
Margo Wood|⏎|Number Of Pages||408|⏎Write a review⏎Note: HTML is not translated!⏎Rating: Bad Good⏎Enter the code in the box below:<|endoftext|>Everyday researchers around the world look for new therapies in
-
specifically occurred in the older days when the trading was done via wind-powered ships. Nowadays, this scenario has become obsolete.⏎Below mentioned are the facts to be kept in mind regarding re export:⏎- The state of items
-
. As a result, they are said to bring relief from illnesses including heart problems, asthma, skin disorder, and arthritis.⏎In the end, you must keep in mind that Ayurvedic<|endoftext|>The number of women who
-
14 x 17 sheet of paper, using the same compressed charcoal. This way you will not get too caught up in detail. Keep these points in mind: Find the basic, simple shapes to describe the character. Find distinctive features
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: in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in in
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
by behaviors including argumentation and silence. These behaviors, in turn, function to reinstate white racial equilibrium and prevent any meaningful cross-racial dialogue. In this in-depth exploration, DiAngelo examines how white fragility develops, how it protects
-
practices⏎This past year, I have had rewarding opportunities to observe traditional food cultures in varied regions of the world. These are:⏎Athabascan Indian in the interior of Alaska (the traditional Tanana Chiefs Conference tribal lands) in July,
-
on either side, down to a massive bend in the Urubamba River. What could have motivated the Incas to undertake such a construction at this remote location in the Andean cordillera? Machu Picchu, built over 7
-
concomitant users of warfarin and a skeletal muscle relaxant and, as pre-specified negative control comparators, among concomitant users of an in
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
- L21:1935 strongly suppresses
- L10:262 strongly suppresses
- L17:706 excites
- L18:2356 excites
- L4:289 excites
- L18:483 excites
- L18:869 suppresses
- L22:535 excites
- L16:1557 excites 'in' meaning 'in this case' or 'in between'
- L16:1807 suppresses 'in' after a motion verb
- L8:351 suppresses 'in' after 'high' or 'dressed'
- L5:490 excites
- L20:1574 excites
- L23:2829 excites
- L3:493 suppresses
- L19:2592 suppresses
- L14:1625 excites the word 'in' after a noun
- L19:704 excites
- L4:1966 suppresses
- L21:1269 suppresses "the" in "on the other hand"
- L16:2120 suppresses
- L14:1122 suppresses
- L20:2850 excites
- L8:3068 suppresses
Feeds later layers
- L26:207 strongly excites
- L27:405 suppresses
- L26:825 suppresses
- L26:584 suppresses
- L27:1621 suppresses
- L27:2429 weakly suppresses
- L26:786 weakly excites
- L27:651 weakly suppresses
- L27:2594 weakly suppresses
- L27:2982 weakly suppresses
- L26:2107 weakly excites
- L26:164 weakly excites
- L26:1662 weakly excites
- L26:938 weakly excites
- L27:142 weakly suppresses
- L26:186 weakly excites
- L27:1426 weakly suppresses
- L27:1795 weakly excites
- L26:2401 weakly suppresses
- L27:2277 weakly suppresses
- L27:2506 weakly suppresses
- L26:82 weakly suppresses
- L27:1522 weakly excites
- L27:2486 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
- ient
- ことです
- ients
- 投标人
- @stop
- ilight
- onto
- Stanton
- görüş
- lsen
Suppresses
- in
- sense
- 日在
- context
- context
- 在
- 而在
- 就是在
- 是在
- 在这种