Neuron Explorer

Qwen3-0.6B

capital N starting an abbreviation

A neuron in layer 25 of Qwen3-0.6B. Its strongest snippets are very consistent: they mostly peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron responds to a capital letter 'N' (or 'H') that starts an abbreviation or name, such as NDR, NCI or N.Y.U. Its opposite side reacts to the 'im' in Cimarron. Turning it down made the model repeat the word 'with' endlessly, and turning it up gave 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 activation was highest. Amber marks positive activation; the outlined token is the peak. Hover a token for its strength.

  1. #1 peak

    study from China, but hope that researchers are on the right track and a solution is in sight gauge, as a virologist Christian Drosten of the NDR Podcast "The Coronavirus-Update" reveals.⏎In the study from Beijing,

  2. #2 peak

    , whilst on its way to meet me in London.⏎Finland is essentially a flat country, slightly mountainous towards the north, but even her highest peak (Haldesjock, in Finnish Lapland) is under 4,0

  3. #3 peak

    ism. Paton also illustrates this theme through the development of several characters in the novel: the literal move of characters such as the pregnant girl to rural life in Ndotsheni represents a change to a greater moral sense, while

  4. #4 peak

    enough clinical verification and validation to know and feel comfortable that what we’re detecting with them is clinically meaningful,” said Lynn Sorbara, Ph.D., of NCI’s Division of Cancer Prevention.⏎Different Tests for Different Tumor Molecules⏎

  5. #5 peak

    to monitor the research on BPA. Glass bottles, however, also leach chemicals, including aluminum, explains a study by Clemens Reimann of the Norges Geologiske Undersøkelse in Norway and colleagues. Aluminum is

  6. #6 peak

    however, was strangely reassuring.⏎“Anxiety disorders are the most common condition in psychiatry,” said Dr. Naomi Simon, professor of psychiatry at N.Y.U. School of Medicine and director of the Anxiety and Complicated Grief

Steering

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.

turned down a lot
Prompt

Here is something I have been thinking about lately: with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with with

Most negative on

Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.

  1. #1 peak

    Digital data sets that describe aquifer characteristics of the alluvial and terrace deposits along the Cimarron River from Freedom to Guthrie in northwestern Oklahoma⏎U.S.

  2. #2 peak

    ette contains digitized aquifer boundaries, maps of hydraulic conductivity, recharge, and ground-water level elevation contours for the alluvial and terrace deposits along the Cimarron River from Freedom to Guthrie in northwestern Oklahoma. Ground water in

  3. #3 peak

    overlie the Permian-age Formations.⏎The aquifer boundaries are from a ground-water modeling report on the alluvial and terrace aquifer along the Cimarron River from Freedom to Guthrie in northwestern Oklahoma and published digital surf

  4. #4 peak

    After a seismic swarm, discussed below , the initial strombolian activity localizes from April 2 on a crack on the southwest flank of Pico do Fogo . A cinder cone called Pico Pequ

Connections

Show on map

Which neurons in earlier layers feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.

Fed by earlier layers

  1. L20:2338 suppresses
  2. L20:3030 suppresses
  3. L19:866 suppresses
  4. L21:665 excites
  5. L19:1437 excites
  6. L23:839 suppresses comma before a trailing 'with' clause
  7. L23:1437 weakly excites
  8. L22:2101 weakly excites
  9. L21:1597 weakly excites
  10. L14:1323 weakly suppresses
  11. L19:1036 weakly suppresses
  12. L12:2994 weakly suppresses
  13. L22:2157 weakly suppresses
  14. L18:2811 weakly suppresses
  15. L10:1481 weakly excites
  16. L17:1044 weakly suppresses
  17. L1:1236 weakly excites
  18. L18:2321 weakly suppresses
  19. L17:1659 weakly suppresses 'as' after 'use' or 'taking' (as a model)
  20. L18:251 weakly suppresses
  21. L19:2292 weakly suppresses
  22. L8:1450 weakly suppresses
  23. L2:1234 weakly excites
  24. L14:450 weakly excites

Feeds later layers

  1. L26:741 strongly suppresses
  2. L27:603 strongly suppresses
  3. L27:1443 suppresses
  4. L27:2153 suppresses
  5. L26:223 excites
  6. L26:973 weakly suppresses
  7. L26:82 weakly suppresses
  8. L26:309 weakly suppresses
  9. L26:2350 weakly suppresses
  10. L27:102 weakly excites
  11. L27:1069 weakly excites
  12. L26:1293 weakly excites
  13. L26:719 weakly suppresses
  14. L26:2058 weakly excites
  15. L27:491 weakly excites
  16. L26:2764 weakly excites
  17. L26:1 weakly excites
  18. L27:1872 weakly excites
  19. L26:1640 weakly excites
  20. L26:1124 weakly excites
  21. L26:1836 weakly excites
  22. L27:74 weakly suppresses
  23. L26:1902 weakly excites
  24. L27:277 weakly suppresses

Activation histogram

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.

1 values between -15.97 and -15.062 values between -15.06 and -14.142 values between -11.40 and -10.494 values between -10.49 and -9.586 values between -9.58 and -8.662 values between -8.66 and -7.754 values between -7.75 and -6.846 values between -6.84 and -5.9212 values between -5.92 and -5.0122 values between -5.01 and -4.1041 values between -4.10 and -3.1874 values between -3.18 and -2.27122 values between -2.27 and -1.36338 values between -1.36 and -0.442664 values between -0.44 and 0.47578 values between 0.47 and 1.3815 values between 1.38 and 2.304 values between 2.30 and 3.215 values between 3.21 and 4.122 values between 4.12 and 5.042 values between 5.95 and 6.861 values between 10.52 and 11.431 values between 19.65 and 20.56offnegativepositiveoften

Logit lens

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 layers instead.

Promotes

  1. rocessing
  2. agas
  3. edes
  4. otyping
  5. ETING
  6. ONSE
  7. yms
  8. istration
  9. eding
  10. utomation

Suppresses

  1. with
  2. with
  3. With
  4. With
  5. _with
  6. →with
  7. WITH
  8. avec
  9. WITH
  10. -with