Neuron Explorer

Qwen3-0.6B

capital 'S' starting a name

A neuron in layer 21 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 a capital S that begins a name or word split into pieces, such as the start of Sputnik, Sandoz, or Snoop. It goes quiet or negative on the letters inside older names like Shakespeare. Turning it up produced rambling text that did not clearly connect to names, so the steering adds little evidence.

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

    , has been unveiled in the Oeiras municipality in the Lisbon metropolitan area, the Russian embassy in Portugal told Sputnik.⏎“We will always remember the names of the first daredevils who

  2. #2 peak

    , Hollingshead came on a mission of hallucinogenic proselytisation.⏎ALBERT HOFMANN AND LSD⏎Swiss research chemist for Sandoz⏎First created LSD in 1938⏎Accidentally absorbed

  3. #3 peak

    % b. If a remainder is ever zero, it means that b⏎divides evenly into a, or a is divisible by b.⏎Try this (from Salkind, 1966): Use a for-loop and an if

  4. #4 peak

    uffs that fueled American fears about a “bomber gap” and later a “missile gap.” On October 4, 1957, Sputnik, the Soviet satellite that was the first to orbit Earth, shocked Americans

  5. #5 peak

    's not only a lodestar of hip-hop, but a global cultural phenomenon. Recent attempts have even been made to resurrect him: He performed in CGI form with Snoop Dogg and Dr. Dre in 2012 and conversed

  6. #6 peak

    about the research than about the paper. Find Scholarly Journals - Writing Your Research Papers - SJSU ... Scholarly articles are kept in databases, usually by discipline or subject

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: it's something to go with a project. A project is a short-song. So, if a project is something to go with a project, then a project is something to go with a project. So, it's a recursive self

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

    THE SURVIVAL of the UNLIKE.⏎Sir M. FOSTER, K.C. B., M. D., and LEWIS E. SHOKÉ,⏎LESSONS IN ELEMENTARY PHY-⏎PHYSIOLOGY. By Sir

  2. #2 peak

    L. partire, partiri, to divide, fr. pars part.⏎1. To part; to divide; to separate. [Obs.] --Shak.⏎2. To go forth or away; to quit, leave, or separate

  3. #3 peak

    x. 30.⏎Ere thou from hence depart. --Milton.⏎He which hath no stomach to this fight,⏎Let him depart. --Shak.⏎3. To forsake; to abandon; to desist or deviate

  4. #4 peak

    where it grows as a perennial shrub. It is now grown commercially in Japan, but because of our cooler climate it’s strictly an ichinensei shok

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. L18:81 strongly suppresses
  2. L20:287 strongly suppresses
  3. L5:204 strongly excites
  4. L10:233 strongly suppresses
  5. L19:594 strongly suppresses
  6. L17:515 strongly suppresses
  7. L17:337 strongly excites
  8. L2:1611 strongly suppresses
  9. L1:5 suppresses
  10. L0:162 excites
  11. L7:287 excites
  12. L2:101 excites
  13. L16:249 suppresses
  14. L17:710 excites
  15. L0:295 excites
  16. L15:89 excites
  17. L16:827 suppresses
  18. L7:300 suppresses
  19. L18:569 excites
  20. L15:114 excites
  21. L15:115 excites
  22. L4:847 suppresses
  23. L16:113 weakly suppresses
  24. L15:82 weakly suppresses

Feeds later layers

  1. L27:2625 excites
  2. L22:1434 excites
  3. L27:2767 excites
  4. L22:666 excites
  5. L26:2636 excites
  6. L27:2045 excites
  7. L24:696 excites
  8. L27:2996 excites
  9. L27:2050 excites
  10. L26:2035 excites
  11. L27:1918 excites
  12. L26:2497 excites
  13. L24:502 suppresses
  14. L27:2225 excites
  15. L27:1871 excites
  16. L26:3004 excites
  17. L26:1555 excites
  18. L26:2437 excites
  19. L27:2425 excites
  20. L26:1151 excites
  21. L26:2376 excites
  22. L27:369 suppresses
  23. L24:431 suppresses
  24. L26:1430 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 -32.56 and -29.752 values between -21.32 and -18.513 values between -15.70 and -12.896 values between -10.08 and -7.2619 values between -7.26 and -4.4552 values between -4.45 and -1.643749 values between -1.64 and 1.1746 values between 1.17 and 3.9811 values between 3.98 and 6.792 values between 6.79 and 9.606 values between 9.60 and 12.411 values between 12.41 and 15.221 values between 15.22 and 18.031 values between 18.03 and 20.852 values between 26.47 and 29.281 values between 29.28 and 32.091 values between 46.14 and 48.961 values between 60.20 and 63.012 values between 77.06 and 79.88offnegativepositiveoften

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. ufficient
  2. atisfied
  3. velte
  4. 背后
  5. creenshot
  6. AVED
  7. chematic
  8. MOOTH
  9. pecially
  10. ensitive

Suppresses

  1. stations
  2. 纩
  3. FLICT
  4. swith
  5. scrim
  6. $s
  7. illed
  8. 沙
  9. tsky
  10. tes