Neuron Explorer

Qwen3-0.6B

dash opening a bullet list item

A neuron in layer 1 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 reacts to the dash at the start of a bullet point in a list, especially right after a line ending in a colon like "includes:". Turning it up or down gave text that did not clearly relate to lists, so the steering results do not explain it.

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

    and dimensions of a structure.⏎Important Steps In the Scanning Process⏎The important steps in the laser scanning process are :⏎- Data collection with a 3D laser scanner⏎- Backing up this data

  2. #2 peak

    is, beyond the particular commands and syntax of a specific language) and develop an understanding of the general principles and characteristics of programming and programming languages. This includes:⏎- algorithms and problem solving with an emphasis on fundamental data structures⏎- object-oriented programming

  3. #3 peak

    in several step using a sequencing algorithm: An algorithm consisting of instructions that are carried out (performed) one after another.⏎We will use the following steps:⏎- Convert all measurement to the same unit (e.g. mm)⏎- Calculate the

  4. #4 peak

    them in sketching. Before you start creating functions in Arduino, understanding facts about functions is vital. Some of the essential facts you will need to note includes:⏎- Every function should have a unique name⏎- A function name must be followed by

  5. #5 peak

    setup () and a loop (). The rest of the functions have to be created outside the bracket with these two functions. The best syntax used in defining function is:⏎- Return type. This is the type of value that is returned by the function and

  6. #6 peak

    part of your business offering, multi-factor authentication is a must. There are many ways to set up multi-factor authentication, but some popular, secure methods include:⏎- Email – let the user receive a verification code or password reset via email⏎-

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 have been thinking about lately, and I have been thinking about lately. So, I have been thinking about lately. So, I have been thinking about lately. So, I have been thinking about lately. So, I have been thinking

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

    Mail newspaper offered a £10,000 prize<|endoftext|>ATLANTA —

  2. #2 peak

    He also received tribunician power and was<|endoftext|>The CDC advises that the COVID-19 vaccination will help protect you from getting COVID-19. In addition, CDC guidance emphasizes proper infection prevention protocols and the proper use of isolation procedures.

  3. #3 peak

    ) for the <codeBase> element, which specifies where the runtime can locate an assembly (for more information, see Specifying an Assembly's Location).⏎<codeBase version="188.8.131.52"

  4. #4 peak

    supply chain may charge more due to changes in petrol prices and the crude price. Due to the state taxes every region price is set accordingly<|endoftext|>Links / Pages:⏎Page at a Gl

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. L0:1301 suppresses
  2. L0:1224 suppresses
  3. L0:178 suppresses
  4. L0:799 weakly suppresses
  5. L0:747 weakly suppresses
  6. L0:732 weakly excites
  7. L0:639 weakly excites
  8. L0:2409 weakly excites
  9. L0:508 weakly suppresses
  10. L0:999 weakly excites
  11. L0:1936 weakly suppresses
  12. L0:1783 weakly excites
  13. L0:9 weakly suppresses
  14. L0:1 weakly suppresses
  15. L0:1507 weakly excites
  16. L0:828 weakly excites
  17. L0:527 weakly excites
  18. L0:202 weakly excites start of a new document
  19. L0:1049 weakly suppresses
  20. L0:1404 weakly excites
  21. L0:2599 weakly suppresses
  22. L0:585 weakly excites
  23. L0:2895 weakly suppresses
  24. L0:2282 weakly excites

Feeds later layers

  1. L23:976 weakly excites
  2. L7:2477 weakly suppresses
  3. L20:368 weakly excites
  4. L17:1109 weakly excites
  5. L11:2533 weakly suppresses
  6. L18:2595 weakly excites 'only' in 'not only ... but also'
  7. L5:3033 weakly suppresses
  8. L25:937 weakly excites
  9. L17:2491 weakly excites
  10. L7:2113 weakly suppresses
  11. L10:982 weakly excites
  12. L9:693 weakly suppresses
  13. L19:1116 weakly excites
  14. L19:434 weakly excites
  15. L16:960 weakly suppresses the word "some" mid-sentence
  16. L4:1606 weakly suppresses
  17. L14:323 weakly suppresses
  18. L17:2306 weakly excites
  19. L8:188 weakly excites
  20. L17:1304 weakly excites
  21. L12:1687 weakly suppresses
  22. L17:561 weakly excites
  23. L14:1660 weakly suppresses
  24. L20:1067 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.

5 values between -0.38 and -0.2822 values between -0.28 and -0.19151 values between -0.19 and -0.101893 values between -0.10 and -0.001513 values between -0.00 and 0.09243 values between 0.09 and 0.1955 values between 0.19 and 0.2813 values between 0.28 and 0.381 values between 1.51 and 1.611 values between 1.70 and 1.801 values between 1.80 and 1.892 values between 1.99 and 2.082 values between 2.08 and 2.181 values between 2.18 and 2.271 values between 2.55 and 2.651 values between 2.93 and 3.031 values between 3.12 and 3.222 values between 3.31 and 3.40offnegativepositiveoften

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. ��
  2. ugal
  3. ��
  4. ,default
  5. -found
  6. peated
  7. BSP
  8. BED
  9. ��
  10. Ethi

Suppresses

  1. rometer
  2. itness
  3. ija
  4. لها
  5. COVER
  6. と言い
  7. 还包括
  8. tong
  9. 对她
  10. ium