Neuron Explorer

Qwen3-0.6B

'rem' at the start of words

A neuron in layer 26 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 fires on the syllable 'rem' at the start of a word, like remodeled, remembrance, remnant and remission. When turned up, the model got stuck repeating a single word, so the steering shows nothing clear about meaning.

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

    destroyed by fire in October 1863.⏎For the next 30 years the court moved into a number of temporary quarters: a school, a remodeled jail, a two-story building on the west side of the square and a

  2. #2 peak

    the first fifteen verses of Psalms 105:1-45. This is a general ascription of praise which merges into a call to remembrance of the works of God, and of His government covenant with the people.

  3. #3 peak

    , White Dwarfs or Pulsars?⏎A personal account by Jocelyn Bell Burnell on the discovery of pulsars.⏎Article from Wikipedia on this remnant that can result from the gravitational collapse of a massive star.⏎Overview of what

  4. #4 peak

    1 poster designed by innovative graphic artist Nick Hayes, which served as a step-by-step guide to researching potential Passchendaele graves and lots of inspiration for remembrance activities within groups and local community, including wreath making and playing the

  5. #5 peak

    /proclivities, etc. Psychometric profiles of civilian<|endoftext|>Recently we were called out to investigate a remnant shipwreck that had materialized on the shoreline near Georgetown, Prince Edward

  6. #6 peak

    law purged with blood, and without shedding of blood is no remission” (Hebrews 9:22). To purge is to

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: my love for plumskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskinskins

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

    lives on earth.⏎Microbiology is the study of the living organisms which are small and cannot be seen with the naked eye. Some of the microorganisms include viruses, fungi, bacteria, protozoa and algae. Microbiology helps in

  2. #2 peak

    point (2, 4)?⏎November 18, 2010⏎biology / plasmolysis of human red blood cells⏎Plasmolysis of human red blood cells would occur

  3. #3 peak

    ers ampicillin resistance to bacterial hosts.⏎- Neo gene from Tn5, which confers antibiotic resistance to kanamycin in bacteria and to geneticin in eukaryote cells.⏎- Mutant FabI gene (mFabI

  4. #4 peak

    —and tasty!—way to make the parts of the cell more concrete for students is to create edible cells. Using materials such as Jell-O® (to represent cytoplasm) and sprinkles (to represent ribosomes) will not only

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. L22:2646 excites
  2. L24:105 weakly suppresses
  3. L18:541 weakly suppresses
  4. L22:1640 weakly excites
  5. L17:253 weakly suppresses
  6. L3:1240 weakly suppresses
  7. L16:3008 weakly excites
  8. L22:2518 weakly suppresses
  9. L0:1849 weakly excites
  10. L16:60 weakly suppresses
  11. L6:2904 weakly suppresses
  12. L0:1365 weakly suppresses
  13. L24:2908 weakly suppresses
  14. L20:1852 weakly suppresses
  15. L3:1325 weakly excites
  16. L6:2953 weakly excites
  17. L8:3018 weakly excites
  18. L5:2060 weakly excites
  19. L9:31 weakly suppresses
  20. L1:2432 weakly suppresses
  21. L6:1759 weakly suppresses
  22. L10:612 weakly suppresses full stop ending boilerplate or a title
  23. L20:1886 weakly suppresses
  24. L16:44 weakly excites

Feeds later layers

  1. L27:2842 weakly excites
  2. L27:365 weakly excites
  3. L27:2939 weakly suppresses
  4. L27:1322 weakly excites
  5. L27:492 weakly excites
  6. L27:1662 weakly excites
  7. L27:2390 weakly excites
  8. L27:2331 weakly excites
  9. L27:1541 weakly suppresses
  10. L27:2376 weakly suppresses capital D starting a split word
  11. L27:155 weakly excites
  12. L27:2992 weakly suppresses
  13. L27:2577 weakly excites
  14. L27:782 weakly excites
  15. L27:2809 weakly suppresses
  16. L27:976 weakly excites
  17. L27:1573 weakly excites
  18. L27:1754 weakly excites
  19. L27:3055 weakly suppresses lowercase 'l' starting a word
  20. L27:489 weakly excites
  21. L27:1770 weakly suppresses
  22. L27:2885 weakly excites
  23. L27:2528 weakly excites
  24. L27:496 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.

1 values between -30.11 and -28.712 values between -28.71 and -27.321 values between -20.33 and -18.941 values between -16.15 and -14.751 values between -14.75 and -13.351 values between -13.35 and -11.961 values between -11.96 and -10.565 values between -10.56 and -9.166 values between -9.16 and -7.778 values between -7.77 and -6.3710 values between -6.37 and -4.9714 values between -4.97 and -3.5826 values between -3.58 and -2.1887 values between -2.18 and -0.783553 values between -0.78 and 0.61145 values between 0.61 and 2.0114 values between 2.01 and 3.4115 values between 3.41 and 4.806 values between 4.80 and 6.205 values between 6.20 and 7.601 values between 7.60 and 8.992 values between 11.79 and 13.181 values between 14.58 and 15.981 values between 18.77 and 20.161 values between 24.35 and 25.75offnegativepositiveoften

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. 轵
  2. men
  3. ijkl
  4. dbe
  5. rze
  6. LinkedIn
  7. физ
  8. Shutterstock
  9. _cmos
  10. ukt

Suppresses

  1. .Extension
  2. ương
  3. trä
  4. perimental
  5. conserv
  6. 즈
  7. 袍
  8. angs
  9. まと
  10. kins