Neuron Explorer

Qwen3-0.6B

"discrete" and "dispersive" in technical text

A neuron in layer 3 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron lights up on technical words that start with "discrete", "dispersive" or "dispersed", such as "discrete math" and "energy dispersive X-ray". It turns negative on the name "Curie" as in "Marie Curie". Turning it up did not produce a clear visible theme in the model's writing.

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

    to Graphs⏎A formal introduction to graphs may be found in most discrete math textbooks. The purpose of this section is not to provide such an introduction.

  2. #2 peak

    analyses of experiments to improve the understanding of the phenomenology.⏎The Large Atomic/Molecular Massively Parallel Simulator (LAMMPS) is a Discrete Element Method (DEM) code developed at Sandia National Laboratories. The code is

  3. #3 peak

    and innovation. Analysis of the economic impact of Waxman-Markey projects that by 2035 the bill will<|endoftext|>Electron Microscopy with Energy Dispersive X-ray Spectroscopy⏎The ability for an aerosol particle to accom

  4. #4 peak

    scanning (SEM) or transmission electron microscopy (TEM) coupled with energy dispersive x-ray spectroscopy (EDX). This technique can provide not only information on

  5. #5 peak

    Students are expected to develop facility with mathematical language and symbols.⏎Because of the central and prerequisite position of discrete mathematical ideas within the field of computer science, study of discrete math is required. If a student does not have prior learning in discrete mathematics,

  6. #6 peak

    - The terms of an arithmetic sequence are said to grow “linearly.” The terms of a geometric sequence are said to grow “exponentially.”⏎- The discrete model of linear growth is ; where is the first term and the common difference.

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: to be a part of the world, to be a part of the family. So, I can be the best I can do, and I can be the best I can say, and I can be the best I can do. I

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

    these 2 projects forms a solid basis to support policy initiatives in Flanders as well as several scientific projects (e.g. Leuven.cool or Curieuzeneuzen). The statistics derived from the garden map have already been designated

  2. #2 peak

    age of 18 living with them, 65.2% were married<|endoftext|>Wellcome to Cornucopia!⏎Cornucopia is a Marie Curie Initial Training Network that started in february 2011, funded by

  3. #3 peak

    Their Life Work Ended Their Lives⏎Their work was for the benefit of the world, but they died as a direct or indirect result of their inventions. Marie Curie won

  4. #4 peak

    Biquard, translated by Geoffrey Strachan (1966). Frédéric Joliot-Curie: the man and his theories. Eriksson. p. 129

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. L0:1224 weakly suppresses
  2. L1:1504 weakly excites
  3. L0:178 weakly suppresses
  4. L0:1301 weakly suppresses
  5. L2:122 weakly excites
  6. L2:1657 weakly excites
  7. L0:2850 weakly excites
  8. L2:2081 weakly suppresses
  9. L2:75 weakly suppresses
  10. L2:70 weakly excites
  11. L2:3003 weakly excites
  12. L0:419 weakly suppresses
  13. L0:17 weakly suppresses
  14. L2:2136 weakly suppresses
  15. L2:2390 weakly excites
  16. L2:2767 weakly excites
  17. L0:2007 weakly suppresses 'key' after 'The'
  18. L0:3006 weakly excites
  19. L1:92 weakly excites
  20. L2:268 weakly excites
  21. L2:1673 weakly excites
  22. L0:2262 weakly excites
  23. L1:433 weakly excites
  24. L0:1887 weakly suppresses

Feeds later layers

  1. L4:3030 excites
  2. L4:1458 suppresses
  3. L15:1010 weakly excites year digits inside 20xx dates
  4. L5:1827 weakly excites the 'D' in 'vitamin D'
  5. L4:2854 weakly excites
  6. L14:1458 weakly suppresses
  7. L23:1141 weakly suppresses
  8. L18:1562 weakly excites
  9. L6:1108 weakly suppresses
  10. L4:2825 weakly suppresses "to" after 'for X' infinitives
  11. L17:558 weakly suppresses
  12. L11:1361 weakly suppresses
  13. L22:1053 weakly suppresses
  14. L6:520 weakly suppresses
  15. L12:2634 weakly suppresses
  16. L15:2029 weakly excites
  17. L5:2724 weakly suppresses
  18. L23:798 weakly excites
  19. L4:794 weakly suppresses
  20. L15:1407 weakly excites
  21. L5:2614 weakly suppresses
  22. L17:1870 weakly suppresses
  23. L18:1887 weakly suppresses
  24. L9:1812 weakly excites

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.

2 values between -0.67 and -0.5111 values between -0.51 and -0.35151 values between -0.35 and -0.191106 values between -0.19 and -0.032341 values between -0.03 and 0.13257 values between 0.13 and 0.2936 values between 0.29 and 0.451 values between 1.08 and 1.241 values between 1.24 and 1.401 values between 1.72 and 1.881 values between 5.53 and 5.69offnegativepositiveoften

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. 流产
  2. addafi
  3. kul
  4. 筹码
  5. awl
  6. 好吧
  7. Explosion
  8. lier
  9. 要害
  10. ضة

Suppresses

  1. alem
  2. 命名为
  3. cro
  4. enames
  5. erged
  6. ahead
  7. Tul
  8. plement
  9. plemented
  10. 命名