Neuron Explorer

Qwen3-0.6B

joining word in a list

A neuron in layer 10 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 a connecting word such as 'and' or 'including' that comes right after a list of names or items. When it was turned up, the written text drifted into unrelated topics, so the steering results do not show a clear effect.

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

    International Space Station (ISS), making it the largest science lesson streamed live for the world to watch on YouTube. A prestigious panel of scientists, astronauts and educators which include His Royal Highness Prince Sultan bin Salman bin Abdul-Aziz Al Saud,

  2. #2 peak

    - Identify points where team members make decisions⏎- Note whether all members of the team are involved.<|endoftext|>The acronym SWOT stands for strengths, weaknesses, opportunities and threats. Organizations of all sizes use SWOT analysis to assess the effectiveness of their

  3. #3 peak

    the same as the height of landing, how fast was the athlete going in the horizontal and vertical directions right before landing? 2. Relevant equations vx = vcos(thata) vy

  4. #4 peak

    data structures that post-quantum algorithms will likely require.⏎We're indebted to Erdem Alkim, Léo Ducas, Thomas Pöppelmann and Peter Schwabe, the researchers who developed “New Hope”,<|endoftext|>Despite the continued

  5. #5 peak

    % of IBS sufferers have been helped with a low FODMAP diet.⏎FODMAPs stands for Fermentable, Oligo-, Di- and Mono-saccharides And Polyols. These are a type of carbohydrates that

  6. #6 peak

    experts say.⏎Prevent Blindness America urges consumers to attend only authorized public fireworks displays conducted by licensed operators. Individuals should not buy, use or store fireworks, including sparklers,

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: I need to be... what? I need to be... I need to be... what... what? what... what? Okay, so the user is looking for a lot here, I need to be... I need to be... 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

    rescue and medical personnel; and U.S. Navy and Coast Guard ships are deployed to support launch contingencies and astronaut recovery. Additionally, the Navy provides a KC-130 tanker for helicopter air refueling, E-2C aircraft

  2. #2 peak

    wolves.⏎tAsAM ruditashabdena gavAM haMbhAraveNa cha |⏎vrajasyotthApanaM chakrurghoShavR^iddhAH samAgatAH ||

  3. #3 peak

    01), systematically split into 6 supergroups, lettered from A to F (Lo et al., 2002). Those in supergroups C and D have coevolved with their nematod

  4. #4 peak

    What is the difference between 3-axis and 5-axis?⏎Three-axis machining is processed by linear feed axes X, Y, and Z.⏎5 axis cnc machining consists of linear interpolation movement of feed axes X, Y, Z and

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. L5:149 strongly suppresses
  2. L2:1661 strongly excites
  3. L1:11 strongly suppresses
  4. L7:342 strongly excites
  5. L0:1010 excites
  6. L4:969 excites the letter C in vitamin C
  7. L1:340 excites
  8. L0:2910 excites
  9. L9:48 suppresses
  10. L3:516 excites
  11. L5:143 excites
  12. L7:287 suppresses
  13. L7:674 suppresses
  14. L2:114 suppresses
  15. L0:2939 weakly suppresses
  16. L0:1996 weakly excites
  17. L4:334 weakly excites
  18. L0:1388 weakly excites
  19. L6:944 weakly excites
  20. L1:50 weakly excites
  21. L0:2231 weakly excites
  22. L5:1712 weakly excites
  23. L0:2303 weakly suppresses
  24. L4:1860 weakly excites

Feeds later layers

  1. L18:88 strongly excites
  2. L16:374 strongly excites
  3. L17:241 strongly excites
  4. L20:314 strongly excites
  5. L19:381 strongly excites word-start 'acc' fragments
  6. L21:320 strongly excites
  7. L18:440 strongly excites
  8. L19:563 strongly excites
  9. L18:541 excites
  10. L21:493 excites
  11. L18:936 excites
  12. L22:607 excites
  13. L22:1155 excites
  14. L20:1486 excites
  15. L17:1216 excites
  16. L21:979 excites
  17. L22:2365 excites
  18. L18:816 excites
  19. L21:590 excites
  20. L27:2597 excites
  21. L18:1493 excites
  22. L26:851 excites
  23. L23:2641 excites
  24. L23:2107 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.

1 values between -2.18 and -2.081 values between -1.59 and -1.491 values between -1.39 and -1.293 values between -1.29 and -1.192 values between -1.09 and -0.993 values between -0.89 and -0.7910 values between -0.79 and -0.6913 values between -0.69 and -0.5912 values between -0.59 and -0.4954 values between -0.49 and -0.3998 values between -0.39 and -0.29218 values between -0.29 and -0.19325 values between -0.19 and -0.09539 values between -0.09 and 0.01716 values between 0.01 and 0.11600 values between 0.11 and 0.21467 values between 0.21 and 0.31312 values between 0.31 and 0.41216 values between 0.41 and 0.51131 values between 0.51 and 0.6169 values between 0.61 and 0.7140 values between 0.71 and 0.8127 values between 0.81 and 0.9118 values between 0.91 and 1.0112 values between 1.01 and 1.116 values between 1.11 and 1.214 values between 1.21 and 1.312 values between 1.31 and 1.414 values between 1.41 and 1.511 values between 1.51 and 1.612 values between 1.61 and 1.711 values between 1.71 and 1.81offnegativepositiveoften

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. esium
  2. oftware
  3. apult
  4. oufl
  5. financed
  6. .PerformLayout
  7. 社
  8. tratt
  9. 郃
  10. Yugosl

Suppresses

  1. しい
  2. ower
  3. asted
  4. raham
  5. ixed
  6. emploi
  7. teacher
  8. ża
  9. enticate
  10. 迁