Neuron Explorer

Qwen3-0.6B

end of document token after a finished passage

A neuron in layer 12 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 mostly lights up at the marker that separates one document from the next, especially after a passage has finished its last sentence. It also fires on a few cut-off words at the end of a text. Turning it up made the writing repetitive and dull.

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

    was packed with (hidden) learning.⏎Here's how the completed assignment looks:⏎Six little math problems, scissors and glue, and a whole lot of learning.<|endoftext|>Popular Inventions

  2. #2 peak

    to have no cost. An example of free is a place where citizens may speak and act as they please, as long as they respect this right of others.<|endoftext|>Bifunctionals show that sometimes simply tethering two useful molecules together unlocks some

  3. #3 peak

    Exposure Attention Interpretation Exposure Exposure highlights the stimulus to which a consumer is exposed to. Exposure can be either deliberate or random. ...⏎Not exactly what you need?<|endoftext|>Allegany County’s

  4. #4 peak

    uous clothing; equipment; outfit.⏎- To cover with a caparison, as a horse.⏎- To dress sumptuously; adorn with rich dress.<|endoftext|>The idea of transgenerational epigenetic inheritance of acquired behaviors is in the

  5. #5 peak

    least 12 missions to carry cargo to and from the International Space Station as part of the Commercial Resupply Services (CRS) contract for NASA.⏎For<|endoftext|>|1st person||me me||nos us|⏎|2nd person||

  6. #6 peak

    the use of technologies in the process of training.⏎the task like that was defined as a task of a multicriteria choice under conditions of uncertainty.⏎To compare the<|endoftext|>A Friday story in National Geographic reports that NASA has found a way to restart 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 meaning of the word s I have to put into the kind of things I have to put into the kind of things I had to put into the kind of things I had to put into the kind of things. So I need to put

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

    . "Sir G. C. Whitlock (During the 1834 Coorg Campaign)". www.archerfamily.org.uk. Archer Family. Archived from the original on 13 November 2013. Retrieved 1

  2. #2 peak

    can tolerate both full sun and partial shade.<|endoftext|>Copyright © University of Cambridge. All rights reserved.⏎'The Solid' printed from http://nrich.maths.org/⏎is one of a series of problems designed to develop learners' team⏎working

  3. #3 peak

    uggle for Power in the Ussr. StudyMode.com. Retrieved 05, 2012, from http://www.studymode.com/essays/The-Struggle-For-Power-In-The-99

  4. #4 peak

    instruction in recognizing alternative interpretations in their writing. (A sample story and four figures of data are included.) (RS)⏎Publication Type: Speeches/Meeting Papers; Reports - Research⏎Education Level: N/A⏎Authoring Institution: N/A

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. L10:58 excites
  2. L8:1060 suppresses
  3. L10:1144 suppresses sentence openers after a break
  4. L10:196 suppresses
  5. L0:2172 weakly suppresses
  6. L10:553 weakly suppresses list-style essay phrasing, weak
  7. L5:323 weakly excites
  8. L11:29 weakly excites
  9. L4:21 weakly excites
  10. L4:1564 weakly suppresses
  11. L3:1596 weakly suppresses
  12. L10:518 weakly excites
  13. L9:248 weakly excites
  14. L8:2010 weakly suppresses
  15. L10:351 weakly excites
  16. L9:948 weakly suppresses
  17. L9:1573 weakly suppresses
  18. L4:2052 weakly excites
  19. L5:1770 weakly suppresses
  20. L0:754 weakly suppresses
  21. L0:2363 weakly suppresses
  22. L3:2218 weakly suppresses
  23. L3:272 weakly excites
  24. L4:623 weakly suppresses

Feeds later layers

  1. L14:2288 suppresses
  2. L18:2100 excites
  3. L13:1001 excites
  4. L13:2328 weakly excites
  5. L13:1513 weakly suppresses
  6. L15:1299 weakly excites
  7. L19:802 weakly excites unclear: word before a company or brand name
  8. L14:586 weakly excites
  9. L13:2766 weakly excites
  10. L25:351 weakly suppresses
  11. L25:239 weakly suppresses
  12. L15:1936 weakly excites
  13. L14:2556 weakly suppresses
  14. L17:892 weakly suppresses
  15. L15:2628 weakly suppresses
  16. L18:1075 weakly suppresses
  17. L18:1136 weakly suppresses
  18. L23:1196 weakly suppresses
  19. L27:1984 weakly excites
  20. L19:2054 weakly suppresses
  21. L21:121 weakly excites
  22. L17:516 weakly excites
  23. L26:320 weakly excites
  24. L21:1856 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.

50 values between -0.54 and -0.40312 values between -0.40 and -0.251177 values between -0.25 and -0.101773 values between -0.10 and 0.05507 values between 0.05 and 0.2074 values between 0.20 and 0.357 values between 0.35 and 0.493 values between 0.49 and 0.641 values between 1.24 and 1.381 values between 1.53 and 1.681 values between 2.42 and 2.571 values between 2.57 and 2.721 values between 5.24 and 5.39offnegativepositiveoften

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. eteria
  2. eson
  3. lesia
  4. esy
  5. roring
  6. ament
  7. esan
  8. OffsetTable
  9. erral
  10. kea

Suppresses

  1. 一流的
  2. 不断的
  3. nice
  4. “⏎⏎
  5. ~
  6. led
  7. nuclei
  8. mond
  9. !”
  10. ~~