Neuron Explorer

Qwen3-0.6B

'official' and government institution words

A neuron in layer 20 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 words like 'official', 'office' and 'government' when they appear in formal, institutional writing. Turning it up pushed the model into repetitive loops about 'power' and 'control', and turning it down gave off-topic 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

    , to retain the full citation information.⏎- This article was written and prepared by U.S. Government employees on official time, and is therefore in the public domain.⏎- This publication may be available

  2. #2 peak

    asuring Housing Affordability⏎U.S. Department of Housing and Urban Development⏎The views expressed in this article are those of the author and do not represent the official positions or policies of

  3. #3 peak

    technical skills are required only in the IT sector, or in office-setting jobs. But that’s not true. Some level of technical knowledge and skills

  4. #4 peak

    Act itself, found in the Statutes at Large, 80 Stat. 383 stated:⏎"Except as otherwise required by statute, matters of official record shall be made available, in accordance with published rule, to persons properly and

  5. #5 peak

    , color, sex, national origin, age, and religion. It guarantees equal opportunity for individuals with disabilities in public accommodations, employment, transportation, State and local government services, and telecommunications.⏎US Dept of Justice ADA website⏎California Department of Development

  6. #6 peak

    be a policy of official discouragement of caste. Many a times, the constitutional courts have asked for the

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: the power over the power over [in the power over] [in the power over] [in the power over] [in the power over] [in the power over] [in the power over] [in the power over]

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

    —they were forced into servitude by circumstances over which they had no control. A quick check of Jesus’ parables shows the relationships between slaves/s

  2. #2 peak

    like something that you talk about when dealing with finances. But this system actually describes how the three branches of government work together to ensure that no one person or group has all of the power.⏎When the Constitution of the United States

  3. #3 peak

    matrices, calculus, binomials, roots, etc.⏎For such sophisticated expressions, most academics, researchers and the like use TeX Equations to have full control over formatting and rendering their equations exactly as they intended

  4. #4 peak

    and Balances⏎The process of creating the Constitution was a real challenge. Everyone involved had concerns about how the government would be run and what person or group would have power. One person who heavily influenced the idea of checks and balances was the Baron

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. L17:2465 excites
  2. L9:1519 excites
  3. L8:3029 excites
  4. L8:2673 suppresses
  5. L18:2581 excites
  6. L12:898 suppresses
  7. L19:2396 suppresses
  8. L12:2471 suppresses
  9. L12:592 weakly suppresses
  10. L18:3049 weakly suppresses
  11. L5:2480 weakly excites
  12. L18:2198 weakly suppresses
  13. L18:2570 weakly suppresses
  14. L18:2510 weakly excites
  15. L4:3023 weakly excites
  16. L17:2888 weakly excites
  17. L18:2053 weakly excites
  18. L18:1649 weakly suppresses
  19. L12:2953 weakly suppresses
  20. L18:2633 weakly suppresses
  21. L15:1644 weakly excites
  22. L17:1485 weakly suppresses
  23. L8:2699 weakly excites
  24. L18:1286 weakly suppresses

Feeds later layers

  1. L24:2746 suppresses
  2. L25:1872 suppresses
  3. L25:2403 suppresses
  4. L27:1416 suppresses
  5. L25:2072 suppresses
  6. L22:2290 suppresses
  7. L24:2697 suppresses
  8. L21:2448 suppresses "a" after give or get verbs
  9. L21:2804 suppresses
  10. L22:2861 suppresses
  11. L24:1250 suppresses mixed: 'New' in place names, square roots
  12. L23:2368 suppresses
  13. L22:2074 weakly suppresses
  14. L24:2954 weakly suppresses
  15. L27:1464 weakly suppresses
  16. L27:1468 weakly suppresses
  17. L22:2752 weakly suppresses
  18. L22:1752 weakly suppresses
  19. L22:1104 weakly suppresses
  20. L27:2501 weakly suppresses
  21. L26:2515 weakly suppresses
  22. L21:2156 weakly excites
  23. L21:1000 weakly excites
  24. L22:1158 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.

2 values between -12.61 and -12.071 values between -11.53 and -10.981 values between -8.27 and -7.731 values between -7.73 and -7.191 values between -7.19 and -6.653 values between -6.11 and -5.561 values between -5.56 and -5.021 values between -5.02 and -4.481 values between -4.48 and -3.944 values between -3.94 and -3.4010 values between -3.40 and -2.857 values between -2.85 and -2.317 values between -2.31 and -1.7731 values between -1.77 and -1.2356 values between -1.23 and -0.69621 values between -0.69 and -0.142349 values between -0.14 and 0.40723 values between 0.40 and 0.9475 values between 0.94 and 1.486 values between 1.48 and 2.022 values between 2.02 and 2.572 values between 3.11 and 3.652 values between 3.65 and 4.191 values between 8.53 and 9.07offnegativepositiveoften

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. imations
  2. imen
  3. iments
  4. Stranger
  5. iment
  6. ilm
  7. ressed
  8. ем
  9. .netbeans
  10. ethoven

Suppresses

  1. control
  2. Control
  3. 控制
  4. CONTROL
  5. /control
  6. power
  7. control
  8. controle
  9. 掌控
  10. 權