Neuron Explorer

Qwen3-0.6B

comma in a list of paired nouns

A neuron in layer 10 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 a comma placed between two or three items in a list, such as 'storage, management' or 'letters, numbers'. Turning it up produced rambling text with no clear link to the trigger.

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

    capabilities, and why should your business consider using it?⏎Put simply, Terraform is powerful infrastructure as code (IaC) tool that allows you to create, update, and control versioning of cloud infrastructure. With Terraform, you can

  2. #2 peak

    , caricatures and sheet music document opera's dramatic 400 year history. From the evocative to the humorous, these images capture opera's artistic, social, economic and political influence, as well

  3. #3 peak

    中具有较大的作用,使用率高,但是泰国学生经常在学习和使用汉语助动词中产生各种的偏误。本文通过调查法、统计

  4. #4 peak

    file represents a sequence of characters (letters, numbers or special characters) that are stored on a personal computer, accessible only to

  5. #5 peak

    ocean...⏎and hydrologic sciences⏎Hydrology is the study of the movement, distribution, and quality of water on Earth and other planets, including the hydrologic

  6. #6 peak

    improved.⏎What is the Cloud?⏎The cloud, in the context of business operations, refers to a network of remote servers hosted on the internet that enable the storage, management and processing of data and applications. Unlike traditional on-premises infrastructure

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: what is the value of the problem? I am not sure if it's something that is worth considering, and I think it might be something that is worth considering. I need to know if it's something that is worth considering. Let's

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

    several technical challenges; (1) creating a robust shell with adequate wall thickness, (2) how to integrated the speaker grille into the plaster form, and (3) creating a suitable mold for the plaster that could could be re-used and not

  2. #2 peak

    al factors: unbalanced nutrition, excessive exercise and/or trauma⏎Most foals will be born with some degree of limb deviation, mostly

  3. #3 peak

    protect animals there. He began to talk to people about animals’ rights, he sought support for his views among his influential friends, he used his money as well as his diplomacy, and eventually he got some animal protection laws<|endoftext|>Play is one of

  4. #4 peak

    with patients who have speech and/or swallowing problems. Someone who has suffered a stroke definitely needs speech therapy in order to help him feed himself independently without choking as well as to restore his speech function. The speech therapist also works in schools with children who

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. L7:1223 suppresses "and" joining two decades or times
  2. L9:749 excites
  3. L8:1238 suppresses
  4. L8:70 excites 'and' between stacked superlatives
  5. L8:593 suppresses
  6. L9:1802 suppresses
  7. L8:1046 excites
  8. L9:458 suppresses
  9. L7:689 excites
  10. L6:620 suppresses "and" in lists of three or more
  11. L7:519 suppresses
  12. L6:427 suppresses
  13. L3:250 excites
  14. L8:479 excites
  15. L6:584 suppresses
  16. L8:986 suppresses
  17. L7:498 suppresses
  18. L5:1469 suppresses
  19. L9:212 excites
  20. L8:1848 suppresses
  21. L6:1706 suppresses
  22. L9:2726 excites
  23. L8:410 excites
  24. L8:131 suppresses

Feeds later layers

  1. L13:656 excites 'and' or comma between paired items
  2. L11:127 excites
  3. L12:959 excites
  4. L11:251 excites
  5. L16:300 excites commas inside lists of nouns
  6. L13:517 excites comma closing a list item
  7. L13:125 excites
  8. L13:1275 excites
  9. L14:672 excites
  10. L12:1212 excites
  11. L15:395 weakly excites
  12. L12:225 weakly excites
  13. L11:682 weakly excites gap before a second number
  14. L12:1756 weakly excites
  15. L11:1595 weakly excites
  16. L13:346 weakly suppresses
  17. L16:761 weakly excites superlatives like 'largest' and 'first'
  18. L17:1499 weakly excites
  19. L13:933 weakly excites
  20. L12:690 weakly excites
  21. L15:1916 weakly excites
  22. L17:1510 weakly excites
  23. L16:2234 weakly suppresses
  24. L16:1766 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.

1 values between -2.98 and -2.831 values between -2.83 and -2.681 values between -2.52 and -2.371 values between -2.37 and -2.221 values between -2.06 and -1.911 values between -1.91 and -1.763 values between -1.76 and -1.603 values between -1.60 and -1.458 values between -1.45 and -1.308 values between -1.30 and -1.1412 values between -1.14 and -0.9920 values between -0.99 and -0.8324 values between -0.83 and -0.6836 values between -0.68 and -0.5378 values between -0.53 and -0.37309 values between -0.37 and -0.22738 values between -0.22 and -0.072027 values between -0.07 and 0.09414 values between 0.09 and 0.2493 values between 0.24 and 0.4036 values between 0.40 and 0.5515 values between 0.55 and 0.7014 values between 0.70 and 0.8612 values between 0.86 and 1.0113 values between 1.01 and 1.1610 values between 1.16 and 1.3210 values between 1.32 and 1.471 values between 1.47 and 1.623 values between 1.62 and 1.783 values between 1.78 and 1.935 values between 1.93 and 2.091 values between 2.39 and 2.551 values between 2.55 and 2.702 values between 2.70 and 2.851 values between 2.85 and 3.012 values between 3.01 and 3.16offnegativepositiveoften

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. ocate
  2. роли
  3. {name
  4. ={}
  5. ammers
  6. $update
  7. {$
  8. <Role
  9. [var
  10. {*

Suppresses

  1. try
  2. trag
  3. ispiel
  4. swallow
  5. try
  6. swallowed
  7. ynth
  8. swallowing
  9. it
  10. puts