Neuron Explorer

Qwen3-0.6B

opening parenthesis in citations and metadata

A neuron in layer 13 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 lights up on the opening bracket in citation style text, like the bracket after 'Copyright' or after a journal name. Turning it up made the model drift into odd, rambling text with no clear theme.

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

    hoped that this paper is a kickstart that will give a useful input to those interested in Artificial Intelligence in education.⏎|Subjects:||L Education > L Education (General)|⏎|Deposited By:||Zaidatun Tasir

  2. #2 peak

    Valentine, Harry T.;⏎Source: Journal Agricultural, Biological, and Environmental Statistics 10:460-477⏎Publication Series: Scientific Journal (JRNL)⏎Description: Transects consisting of multiple, connected segments with a prescribed

  3. #3 peak

    captive crocodilians Authors⏎|Davis, Lisa - APPLIED BIOSYSTEMS INC|⏎Submitted to: Journal of Experimental Zoology⏎Publication Type: Peer Reviewed Journal⏎Publication Acceptance Date: February 1

  4. #4 peak

    water, were the major factors shaping phytoplankton communities.⏎- Abstract views: 1777⏎- PDF: 842⏎Copyright (c) 2009 Silvia TAVERNINI, Daniele N

  5. #5 peak

    , student academic ability, and students learning in their mother tongue are some factors for consideration. Copyright © 2014 Society for the Scientific Study of Reading (SSSR).⏎|Publication status||Published - Jul

  6. #6 peak

    of creating a “thermostatic mechanism” that would counter-balance the dominant presence of electronic media in the lives of children.⏎How to Cite⏎Copyright (c) 2017 Robert Albrecht⏎Copyrights and publishing

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 problem with the problem that I have to do this in the school that I am doing my job, and I need to help the students in the school, and I am responsible for the students and the school in my job in my work

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

    function multiplied by its conjugate. The statement of the evenness of the porobability

  2. #2 peak

    table with a specific name or index exists.⏎The following example retrieves a single table by name from the DataTableCollection.⏎Private Sub GetTableByName() ' Presuming a DataGrid is displaying more than one table

  3. #3 peak

    conuts and banana leaves<|endoftext|>const void * memchr ( const void * ptr, int value, size_t num );⏎

  4. #4 peak

    code, and humans infer things from the code that will be necessary to work on it.⏎typedef void * model_3d_ref;⏎model_3d_ref load_model_from_disk(const char * absolute_file_path);⏎void draw_model(model_3

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. L11:746 excites
  2. L7:158 excites
  3. L12:1196 suppresses
  4. L11:1024 weakly suppresses
  5. L11:1912 weakly excites words right after an 'if' condition
  6. L12:2762 weakly suppresses
  7. L11:2581 weakly excites
  8. L11:1927 weakly excites
  9. L11:3048 weakly excites
  10. L5:897 weakly suppresses
  11. L11:540 weakly suppresses
  12. L10:1157 weakly suppresses
  13. L3:102 weakly suppresses
  14. L3:488 weakly excites
  15. L3:173 weakly suppresses
  16. L3:1607 weakly excites
  17. L10:296 weakly excites
  18. L6:1036 weakly suppresses
  19. L9:1302 weakly suppresses
  20. L8:344 weakly suppresses
  21. L7:656 weakly suppresses
  22. L12:120 weakly excites
  23. L11:1479 weakly excites
  24. L9:1366 weakly suppresses

Feeds later layers

  1. L18:1501 suppresses
  2. L16:2306 suppresses
  3. L16:1777 suppresses
  4. L15:1857 suppresses
  5. L15:1634 suppresses
  6. L19:369 suppresses
  7. L14:2958 suppresses operators and symbols in code or formulas
  8. L26:1331 weakly suppresses
  9. L25:360 weakly suppresses
  10. L20:758 weakly suppresses
  11. L16:2482 weakly excites
  12. L14:1559 weakly suppresses
  13. L23:224 weakly suppresses
  14. L18:1266 weakly suppresses
  15. L15:2582 weakly excites
  16. L19:900 weakly suppresses opening parenthesis after a term
  17. L16:590 weakly excites
  18. L26:476 weakly excites
  19. L16:2116 weakly suppresses
  20. L16:508 weakly suppresses
  21. L20:2956 weakly suppresses
  22. L14:696 weakly suppresses
  23. L18:2141 weakly excites
  24. L15:2335 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 -2.41 and -2.321 values between -0.98 and -0.891 values between -0.89 and -0.812 values between -0.64 and -0.563 values between -0.56 and -0.475 values between -0.47 and -0.3940 values between -0.39 and -0.30108 values between -0.30 and -0.22358 values between -0.22 and -0.14742 values between -0.14 and -0.051173 values between -0.05 and 0.03859 values between 0.03 and 0.12429 values between 0.12 and 0.20137 values between 0.20 and 0.2835 values between 0.28 and 0.378 values between 0.37 and 0.451 values between 0.45 and 0.542 values between 0.54 and 0.621 values between 0.62 and 0.711 values between 0.71 and 0.791 values between 0.87 and 0.96offnegativepositiveoften

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. =yes
  2. dozen
  3. --; ⏎
  4. --;⏎
  5. .ZERO
  6. .LENGTH
  7. ches
  8. DBus
  9. /Math
  10. ��

Suppresses

  1. lage
  2. 而在
  3. 圭
  4. ícia
  5. ervention
  6. điển
  7. лез
  8. bare
  9. rząd
  10. prática