Neuron Explorer

Qwen3-0.6B

'of' in 'this type of'

A neuron in layer 2 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 responds to the word 'of' when it comes right after 'type', as in 'this type of failure' or 'this type of game'. Turning it up did not produce text that clearly reflected this pattern.

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

    for this type of post-disaster performance, Waffle House has become well prepared for Business Continuity

  2. #2 peak

    outcome. While some areas of their specialty may overlap, each member has a distinct role to play in the patient’s recovery.⏎Here are a few:⏎This type of rehabilitation deals with

  3. #3 peak

    stop it, a valve needs to be closed but the situation is too dangerous to send humans in. At the moment, no type of robot can do this type of job and that is what Kim wants to create - a robotic first-responder -

  4. #4 peak

    Zinc is also susceptible to liquid metal cracking, but only when the temperatures are much higher (750F range). The most recently reported case of this type of failure (August 1998) was on a fleet of school buses

  5. #5 peak

    shoulders are intricately decorated with epaulets bearing silver oak leaves on a blue field, indicating the Infantry branch of the United States' Army. This type of dress coat would likely have been worn for portraits and formal occasions.<|endoftext|>Much can

  6. #6 peak

    from a Chinese speaker.⏎The demo also relies on a new method for processing speech patters, using a system called Deep Neural Networks. The company says this type of machine learning is "patterned after human brain behavior," and means that the researchers

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: - - - - - - - - - - - - - - - - - - - - - - - -

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

    house⏎• Add 3D furnishings such as sofas, chairs, beds, and more⏎• Perform 3D photorealistic rendering using the latest rendering engine from AutoCAD⏎• Make an animated walk-through video from your 3D

  2. #2 peak

    grow in complexity with time.⏎The global pattern to follow is MVC and the layering:⏎(View & Controller)⏎|View||HTML templates usually generated using a template engine|⏎|Input Controller||Filter and validate data|⏎||Application Controller||Application flow

  3. #3 peak

    Tok video-sharing app is a psychographic predictive machine learning engine, whose algorithm can intelligently catalog individual profiles through reactions to the seemingly innocuous

  4. #4 peak

    sub-matrix format.⏎The data file format for nmrPipe files is the serial format. It is possible, by

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. L1:254 suppresses
  2. L1:492 weakly excites
  3. L1:386 weakly suppresses
  4. L1:925 weakly suppresses
  5. L1:266 weakly suppresses
  6. L1:1857 weakly suppresses
  7. L0:1640 weakly excites
  8. L1:2667 weakly suppresses
  9. L1:1529 weakly excites
  10. L0:176 weakly excites
  11. L1:166 weakly suppresses
  12. L0:2024 weakly suppresses
  13. L1:1769 weakly excites
  14. L1:92 weakly excites
  15. L1:2683 weakly excites
  16. L0:25 weakly excites
  17. L0:1301 weakly suppresses
  18. L1:173 weakly excites
  19. L1:87 weakly suppresses
  20. L0:317 weakly excites
  21. L1:1043 weakly excites
  22. L0:178 weakly suppresses
  23. L0:174 weakly suppresses
  24. L1:718 weakly suppresses

Feeds later layers

  1. L20:119 weakly excites
  2. L11:2733 weakly excites
  3. L10:104 weakly excites
  4. L5:1731 weakly suppresses
  5. L21:1294 weakly excites
  6. L7:116 weakly excites
  7. L6:3006 weakly excites
  8. L12:1922 weakly suppresses sentence-ending periods in explanations
  9. L6:502 weakly excites
  10. L7:887 weakly suppresses
  11. L5:679 weakly excites
  12. L7:3 weakly suppresses
  13. L8:1984 weakly excites
  14. L6:1906 weakly excites
  15. L10:2172 weakly excites
  16. L6:1555 weakly suppresses
  17. L15:908 weakly suppresses
  18. L10:465 weakly suppresses
  19. L4:855 weakly suppresses
  20. L7:420 weakly excites
  21. L3:576 weakly suppresses
  22. L10:2848 weakly excites
  23. L15:814 weakly suppresses
  24. L3:995 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 -0.91 and -0.842 values between -0.71 and -0.643 values between -0.64 and -0.571 values between -0.50 and -0.4311 values between -0.43 and -0.3614 values between -0.36 and -0.3036 values between -0.30 and -0.23119 values between -0.23 and -0.16353 values between -0.16 and -0.09877 values between -0.09 and -0.021723 values between -0.02 and 0.05511 values between 0.05 and 0.12179 values between 0.12 and 0.1839 values between 0.18 and 0.2518 values between 0.25 and 0.3212 values between 0.32 and 0.394 values between 0.39 and 0.462 values between 0.46 and 0.531 values between 0.53 and 0.601 values between 0.60 and 0.661 values between 1.76 and 1.83offnegativepositiveoften

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. нию
  2. ража
  3. יים
  4. нии
  5. Tues
  6. елеф
  7. irty
  8. .$.
  9. 斗志
  10. POINTER

Suppresses

  1. local
  2. rogate
  3. -local
  4. local
  5. 您的
  6. 强力
  7. كر
  8. 你的
  9. 顾
  10. Local